October 22, 2014

High coverage genome from 45,000-year old Siberian (Ust'-Ishim)

This is the oldest full genome of a modern human published to date and it also comes from a time (45 thousand years ago) that coincides with the Upper Paleolithic revolution in Eurasia.

45 thousand years ago is probably close to when Eurasians started diverging from each other as they spread in all directions. So, we expect that a human from that time would be "undifferentiated Eurasian" and indeed this seems to be the case.

First the Y-chromosome:
The Y chromosome sequence of the Ust’-Ishim individual is similarly inferred to be ancestral to a group of related Y chromosomes (haplogroup K(xLT)) that occurs across Eurasia today6 (Supplementary Information section 9).
and mtDNA:
The Ust’-Ishim mtDNA sequence falls at the root of a large group of related mtDNAs (the ‘R haplogroup’), which occurs today across Eurasia (Supplementary Information section 8).
It is clear that this was a Eurasian individual:
Based on genotyping data for 87 African and 108 non-African individuals (Supplementary Information section 11), the Ust’-Ishim genome shares more alleles with non-Africans than with sub-Saharan Africans (|Z| = 41–89), consistent with the principal component analysis, mtDNA and Y chromosome results.
It was also more like East Asians than Europeans:
Among the non-Africans, the Ust’-Ishim genome shares more derived alleles with present-day people from East Asia than with present-day Europeans (|Z| = 2.1–6.4).
But, when they compared East Asians with La Brana and MA-1 they didn't see a difference:
However, when an ~8,000-year-old genome from western Europe (La Braña)9 or a 24,000-year-old genome from Siberia (Mal’ta 1)10 were analysed, there is no evidence that the Ust’-Ishim genome shares more derived alleles with present-day East Asians than with these prehistoric individuals (|Z| < 2). This suggests that the population to which the Ust’-Ishim individual belonged diverged from the ancestors of present-day West Eurasian and East Eurasian populations before—or simultaneously with—their divergence from each other. The finding that the Ust’-Ishim individual is equally closely related to present-day Asians and to 8,000- to 24,000-year-old individuals from western Eurasia, but not to present-day Europeans, is compatible with the hypothesis that present-day Europeans derive some of their ancestry from a population that did not participate in the initial dispersals of modern humans into Europe and Asia11.
So it seems that the Ust'-Ishim individual belonged to the same branch as Asians and WHG/ANE and modern Europeans are less like it because they also have "Basal Eurasian" admixture which they inherited via the EEF in the model of Lazaridis et al.

The authors could also get estimates of the mutation rate because this is a 45,000 year old individual that hasn't experienced 45,000 years worth of mutations:
Assuming that this corresponds to the number of mutations that have accumulated over around 45,000 years, we estimate a mutation rate of 0.43 × 10−9 per site per year (95% CI 0.38 × 10−9 to 0.49 × 10−9) that is consistent across all non-African genomes regardless of their coverage (Supplementary Information section 14). This overall rate, as well as the relative rates inferred for different mutational classes (transversions, non-CpG transitions, and CpG transitions), is similar to the rate observed for de novo estimates from human pedigrees (~0.5 × 10−9 per site per year14, 15) and to the direct estimate of branch shortening (Supplementary Information section 10). As discussed elsewhere14, 16, 17, these rates are slower than those estimated using calibrations based on the fossil record and thus suggest older dates for the splits of modern human and archaic populations.
This is a very direct confirmation of the "slow" autosomal rate of ~1.2x10-8 mutations/generation/bp using a technology much different than those used before to estimate this. The slower mutation rate implies that major splits in human history (such as the Out-of-Africa event) took place much earlier than the Upper Paleolithic revolution and the spread of humans across Eurasia. Modern humans probably established an early presence in the Levant/Arabia (consistent with Out-of-Arabia), and invented the Upper Paleolithic-related tools/behaviors there much later, and only then spread across Eurasia.

The authors write:
we estimate that the admixture between the ancestors of the Ust’-Ishim individual and Neanderthals occurred approximately 50,000 to 60,000 years BP, which is close to the time of the major expansion of modern humans out of Africa and the Middle East.
This clinches the hypothesis of Neandertal introgression in Eurasians, as Ust'-Ishim has longer Neandertal segments than modern humans, as one might expect from an individual who experienced this admixture more recently in its evolutionary past than modern humans did. It's probably in the Middle East that the Levantine/Arabian modern humans that expanded Out-of-Africa more than 100 thousand years ago came into contact with Neandertals, admixed with them and later carried this ancestry to the rest of Eurasia. I tend to think that the AMH "colony" was first limited to Arabia and only later (post-70kya) expanded north as the climate deteriorated there. The authors estimate the common ancestor of non-African Y-chromosomes (including E, which is probably a back-migration to Africa) to around 70 thousand years ago which may coincide with the Arabian Exodus event.

Nature 514, 445–449 (23 October 2014) doi:10.1038/nature13810

Genome sequence of a 45,000-year-old modern human from western Siberia

Qiaomei Fu et al.

We present the high-quality genome sequence of a ~45,000-year-old modern human male from Siberia. This individual derives from a population that lived before—or simultaneously with—the separation of the populations in western and eastern Eurasia and carries a similar amount of Neanderthal ancestry as present-day Eurasians. However, the genomic segments of Neanderthal ancestry are substantially longer than those observed in present-day individuals, indicating that Neanderthal gene flow into the ancestors of this individual occurred 7,000–13,000 years before he lived. We estimate an autosomal mutation rate of 0.4 × 10−9 to 0.6 × 10−9 per site per year, a Y chromosomal mutation rate of 0.7 × 10−9 to 0.9 × 10−9 per site per year based on the additional substitutions that have occurred in present-day non-Africans compared to this genome, and a mitochondrial mutation rate of 1.8 × 10−8 to 3.2 × 10−8 per site per year based on the age of the bone.

Link

139 comments:

epoch2013 said...

"This suggests that the population to which the Ust’-Ishim individual belonged diverged from the ancestors of present-day West Eurasian and East Eurasian populations before—or simultaneously with—their divergence from each other. "

So, Am I utterly wrong to think that WHG is simply the westernmost part of a separate group of human beings that MA-1 is the easternmost part of?

David Jacobson said...

The most important fact about this genome is the location it came from. It is on the Northern edge of the plateau region that separates Europe, India, Africa, and China. An human living in this region would certainly already have to deal with a very different climate from Africa and a more severe climate than is typical for Europe. It is certainly a place where evolution towards a Siberian population would be expected. It is also in a region that is a good candidate for the differentiation of East Asian Genomes.
The time required for the evolution and its mechanisms remain unknown. The reality that becomes obvious from inspection of full human genomes is that extended series of so called SNP's are the common element of human genetic variation. That variation is common even just among African populations. I am very skeptical of Svante Paabo's idea that these series inevitably start out big and get smaller with time. But, the kind of analysis done here does not even begin to objectively answer that question.

Lank said...

Other than finally establishing the autosomal mutation rate, this paper is a blatant confirmation of the conventional mtDNA and Y-DNA mutation rates. Both of these give us a date for the common African-Eurasian mtDNA/Y-DNA ancestors around ~70 kya, with Y-DNA CT originating in Africa or Eurasia, and mtDNA L3 probably originating in East Africa.

The autosomal divergence is expected to be greater than the dates derived from uniparental markers, for two reasons. Firstly, there is archaic admixture in Eurasia, and probably Africa as well. Secondly, Africans carry other (likely AMH) Y-DNA and mtDNA lines than just Y-DNA CT and mtDNA L3.

However, the archaeological evidence indicates a modern human presence in the Middle East long before 50-70 thousand years ago (the age of the oldest apparent Eurasian Y-DNA/mtDNA lines). So the possibility of continuity from those early migrants should also be considered to explain the date derived from autosomal divergence. But, the other factors leading to a higher autosomal divergence should not be ignored.

agiering said...

"So it seems that the Ust'-Ishim individual belonged to the same branch as Asians and WHG/ANE and modern Europeans are less like it because they also have "Basal Eurasian" admixture which they inherited via the EEF in the model of Lazaridis et al."

Wouldn't this make Ust'-ishim similar to Amerindians, who also are more similar to East Asians, WHG, and Mal'ta (i.e., "ANE") than they are to Stuttgart (i.e., "EEF")?

My goodness, German is going to have a field day with these findings...

G Horvat said...

Unfortunately, the following statement is incorrect:
"The Ust’-Ishim individual carries the following substitutions that define the R haplogroup (compared to rCRS sequence) : 73G, 263G, 750G, 1438G, 2706G, 3107d, 4769G, 7028T, 8860G, 11719A, 14766T, 15326G..."

These are the original forms and therefore do not signify Haplogroup R.

Kes said...

So, now we know it belongs to K2, which seems to have diversified in and then expanded from SE Asia.

I wonder if Ust-Ishim is a remnant of an earlier expansion from India/Central Asia (of which the SE Asian K2 would eventually become the most successful branch) or part of the later expansion from Wallacea?

It seems a bit early for the latter, but who knows?

terryt said...

"The Y chromosome sequence of the Ust’-Ishim individual is similarly inferred to be ancestral to a group of related Y chromosomes (haplogroup K(xLT)) that occurs across Eurasia today"

That means it is K2, probably specifically K2b2 or P. At 45,000 years that is considerably earlier than dates previously suggested for the appearance of the haplogroup. However it does fit with the age suggested for the crossing of Wallacea. But that means the back migration from there was very rapid.

Mark Moore (Moderator) said...

I am getting conflict reports on this research regarding the significance of longer DNA strands which match up with Neandertal DNA. Some reports say that there are a few really long strands which indicated that this extinct population of humans continued breeding with neandertals. Other reports only mention that the longer segments are because this individual was closer in time to the date where they believe the base Eurasian group of humans mated with neandertals. Is it one or the other? Both?

Also I wonder if they measured for only Neandertal segments or also for Denisovan, since that seems to be the non-human that was in the area at the time?

n/a said...

Called it.

eurologist said...

Fascinating. I had guessed P - but the beginning of P is just a few mutations away.

The confirmation of something like a "Basal European"/ EEF component/ admixture is very interesting. I am still not quite sure where to place them and what haplogroups to associate them with. Can't be I or J on the y-DNA side; can be E but then a large impact on the autosomal side is weird. Perhaps G? But where was this "paradise" in tens of thousands of years of adverse climate in SW Asia?

Since archaeology tells us AMHs migrated both west and east from Siberia, this seems to support (a) that this migration from S/ SE Asia followed a route to the west of the Himalayas, and that it occurred extremely shortly after or at roughly the same time as northward migrations inland and along the coast in E Asia.

Grognard said...

Much, or even most, of the genes that differentiate modern europeans probably got fixed only after farming was discovered.

Kristiina said...

There are people who say that Ust’-Ishim is ancestral to West and East Eurasians but I am wondering if his closest population was really ancestral to any modern population and if Ust'-Ishim is rather an uncle or aunt to these populations because: 1) ”The Ust’-Ishim Y-chromosome carries no additional mutations belonging to any of the sub-haplogroups of K(xLT); however, there are 6 additional mutations that are not observed in the 23 present-day humans to which we compare”; his mtDNA is also unique but it is due only to one specific substitution not seen in any presentday population; 2) in admixture tests there is only an admixture edge between Ust’-Ishim and Altai Neandertals but not with any other population; 3) when you look at Figure S17.1. of the supplementary information (sharing of derived alleles with Neandertals), not any pieces of the biggest Neandertal chunk have been passed on to Han, Karitiana or Australians; instead modern populations tend to have yellow Neandertal color on the same or a close vertical line which is usually not shared with Ust’-Ishim. (However, I do not know how much individual variation there is in modern populations in this respect!)

My guess is that people were severely decimated in North Eurasia during the Ice Age and Ust’-Ishim may belong to one of this extinct populations. It would be easy to conclude that, after the Ice Age, North Eurasia was to a great extent repopulated from the more southern areas.

In the admixture analysis, Ust’-Ishmim’s most important refererence populations are Central & South Asia, Oceania and Southeast Asia in this order. Therefore, I would say that it is probable that his yDNA K and mtDNA R came from India which is in line with the haplotrees.

Before the publication of this paperg, I was hoping to see a genome that would have left traces in Europeans, Native Americans and Northeast Asians but it seems that there is not really (any?) gene flow from Ust'-Ishim to these populations.

capra internetensis said...

Very exciting! Lots of good stuff in the supp info.

Y-DNA TMRCA dates based on this calibration (from what data I could find on the HGDP samples they used):
- the root of their tree should be A1b-P108 at 153 (+-22) kya
- A1b1-L419 (San_A is A1b1a, others A1b1b) - 136 (+-21) kya
- CDEF-M168 (no B) - 72 (+-9) kya
- IJK (no other F) - 54 (+-5) kya
- C-M130 (Dai_B is C2, others C1) - 53 (+-8) kya
- E-M96 (both E1 and E2 samples) - 53 (+-8) kya
- K2 (no LT) - 50 (+-4) kya
- IJ - 49 (+-6) kya

So it looks like C, F, and E all diversified at pretty much exactly the same time (beginning of Upper Paleolithic), followed by major branches shortly afterward (E1, C1, IJ, K), but the breakup of CDEF was considerably earlier. The dates are ~15% younger than those in Scozzari et al, so the split of B from CDEF would probably be around 100 kya by this calibration.

German Dziebel said...

Great study! More about it at http://anthropogenesis.kinshipstudies.org/2014/10/ancient-ust-ishim-dna-as-seen-from-the-americas/.

Gregory76 said...

This article seems to support my belief that male K (ex LT) and female R, and their descendents, migrated to their current homes in Europe, East Eurasia and the Americas by way of Central Asia, and not by way of Southeast Asia.s

Unknown said...

a huge thankyou for the work you do publishing these papers. for those of us without access to scholarly databases, it is such a gift. thanks!

without speculating on the location of introgression, and just focusing on synapomorphy suggested by this paper the divergence tree would be:

basal common ancestor 100k, branching to the african lineage and the eurasian lineage

basal eurasian introgressed lineage 60k, branching to a largely derived european lineage (to account for the lesser neanderthal ancestry) and a less derived eurasian lineage, circa 50k

with eurasaisan lineage branching simaltaneously circa 45k from the group that individual studied in this paper to at least three families, east asian, the 8k, and the 40k. the best explanation would be resource exhaustion with the pioneers uncertain which direction would be the most fruitful.

the split with the african lineage could equally be explained by an archaic modern population branching out from eurasia to africa as branching out from africa.

in any other species the into africa would be the accepted theory because it is the simplest and accounts for all of the current data without contradiction. it also is inline with what is accepted for all other speieces-the least similar to all others is the most derived, not the basal. as would be expected for a lineage that broke off 100k years ago.

Unknown said...

quick ps: even back in college when getting my degree in anthropolgy, my candidate for the common ancestor of human-african apes - orangutans was oreopithicus (or proto-oreopithicus), because it has traits found in all and is in a location much more conducive to such a broad radiation.

chimps, bonobos, and gorillas have more derived traits than any other apes. and the presence of hominids in africa is neither necessary nor sufficient for proving our descent from african hominids, esp. when proto-chimp and proto-gorilla have never been found.

the descendants of an oreopithicus-like, eurasian ape would be more than enough to account for wide soread hominodial traits.

humans have more in common morphologically with monkeys, including our molar pattern, than we do with any of the great apes...

ryukendo kendow said...

@ Dienekes
Didn't the authors say that the Neanderthal admxture into Ust was recent?

Dr. Clyde Winters said...

The discovery of the Ust’Ishim ancient mtDNA has important implications for the spread of AMH in Eurasia. The dating of the remains appear problematic. Although the remains eroded from the Burginca and Zogoozdino layers of Siberia dated between 40-24.5kya, and a radio-carbon date of 41,400 BP, researchers have given a calendar date for the remains of 45,770-44,000BP. The date for the layers where the remains were found and radio-carbon date for the remains suggest the probable calendar date for the remains was 40.5ky, not 45ky. A later date for Ust’Ishim would be place it in conformity with AMH from Tianyuan Cave and Pestera cu Oase between 39.5 kya and 39.1kya; and Kostenki 14 and Kents Cavern 4, which date between 37.8 and 35.4 BP.
The finding that the Ust’Ishim AMH did not carry Basal Eurasian ancestry suggest to me at least, that this population was not related to the AMH associated with the Aurignacian AMH who entered Europe via the Straits of Gibraltar. The finding that Ust’Ishim’s mtDNA was related to R* subclades, and y-chromosome K(xLT) indicate that this individual was a descendant of the original OoA event 60kya who had settled in India. This view is supported by the close relationship of the Ust’Ishim to the Papuans, Australians and Onge, Andaman Islands, India.
In summary Ust’Ishim AMH suggest that this population was descendant from the first OoA exit 60kya, rather than the bearers of the Aurignacian culture. The dating of Ust’Ishim between 45-40 kya does not in anyway force us to predict an OoA event prior to 60kya.

Hector said...

Most previous mutation rates were per generation, not per year, so direct comparison is difficult.

The man has M526 but lacks any further mutation shared by extant populations. "Southeast Asian -ness" of K2b still remains. But I foresee that this blog will be inundated with happy wishful thinkers who want to see otherwise and claim moral victory of some sort.

It will be hilarious.

The author may have made an error in not taking it into the consideration that Malta boy etc. are supposed to be closer to this man than modern populations are to him, all other things being equal. This would mean that Ust Ishim man may be closer to East Eurasians already.

gnomon said...

One expect that the Ust'-Ishim genome should be almost 2 fold less distant to living Africans than East Asians are because he had 45,000 years less time to accumulate distance. But the paper made no mention of this key expectation from the Out of Africa model. Why?

terryt said...

"I am still not quite sure where to place them and what haplogroups to associate them with. Can't be I or J on the y-DNA side"

As Kristiin quoted:

"The Ust’-Ishim Y-chromosome carries no additional mutations belonging to any of the sub-haplogroups of K(xLT); however, there are 6 additional mutations that are not observed in the 23 present-day humans to which we compare'"

His is an extinct line that branched off after LT (K1) formed but before K2 had begun diversifying. Therefore it is not ancestral to P, or to any other surviving haplogroup. Of course that doesn't mean at all that none of its a-DNA survives.

"My guess is that people were severely decimated in North Eurasia during the Ice Age and Ust’-Ishim may belong to one of this extinct populations".

My guess too.

"'Southeast Asian -ness' of K2b still remains. But I foresee that this blog will be inundated with happy wishful thinkers who want to see otherwise and claim moral victory of some sort".

It's already started at other blogs. I know you agree that K2b cannot have originated anywhere other than in SE Asia.

Gregory76 said...

TerryT said
“That means it is K2, probably specifically K2b2 or P. At 45,000 years that is considerably earlier than dates previously suggested for the appearance of the haplogroup. However it does fit with the age suggested for the crossing of Wallacea. But that means the back migration from there was very rapid.”

Yet another reason for doubting that K2 and P migrating to northern and central Eurasia from southeast Asia.

Hector said
“"Southeast Asian -ness" of K2b still remains.”

Merely asserting that claim in the face of this evidence to the contrary does nothing to argue for the claim.


eurologist said...

"My guess is that people were severely decimated in North Eurasia during the Ice Age and Ust’-Ishim may belong to one of this extinct populations. It would be easy to conclude that, after the Ice Age, North Eurasia was to a great extent repopulated from the more southern areas.

Before the publication of this paperg, I was hoping to see a genome that would have left traces in Europeans, Native Americans and Northeast Asians but it seems that there is not really (any?) gene flow from Ust'-Ishim to these populations."


Kristiina,

Yes and no. Archaeological evidence is that Siberia was continuously and almost contiguously populated during most if not all of the ice age.

However, certainly population numbers were small and prone to overwhelming influx (mostly from the immediately adjacent southeast, judging by haplogroups).

But don't forget that autosomal heritage is very different from uniparental haplogroups. Even if there is no direct descendant from the latter perspective, Ust'-Ishim is autosomally extremely close to the common ancestor of all Eurasians and Native Americans.

capra internetensis said...

@Gregory76 and Hector:

we will need to know more about K2 and the age of divergence of its subclades to settle the question of its origins. Clearly K2 had a very old and successful expansion in SE Asia, but as P1 split off very early it is possible that it is a lineage that remained from the initial expansion of K2 in South (or Central) Asia and got lucky 12 ky later.

@gnomon:

the split from Africans probably happened ~100 kya going by the Y-chromosomal data. So he would be about 155 ky from modern Africans, while East Asians would be 200 ky, making him 3/4 as distant. However, I doubt such a simplistic calculation means much. Ust'-Ishim man has more mixture from Neanderthals, a population that has been diverging for 500 ky+, than do living Africans; the Out-of-Africa group presumably split from one subgroup of Africans, not from a representative sample of the African population, which includes much more divergent ancestry; and there has been gene flow into and out the continent since then. The first two factors would tend to make the distance greater, I think, and the latter would make it smaller. (If I understand this stuff correctly.)

If you check the world PCA plot in the supplementary info, you'll see that Ust'-Ishim man is actually projected a little to the African side of South Asians.

Tobus said...

@gnomon: One expect that the Ust'-Ishim genome should be almost 2 fold less distant to living Africans than East Asians are because he had 45,000 years less time to accumulate distance.

Your maths is wrong - assuming a (very late!) 100kya separation date then there is 200ky worth of mutations between modern populations (100ky worth in Africa plus 100ky in Eurasia). In comparison, at 45kya we'd expect Ust'-Ishim to have 155kya worth of mutations... 155 to 200 is not "2 fold less", it's about 25% less. Using an earlier (and much more likely!) separation date reduces this expected difference even more. If you check figures 8.1, 9.1 and 10.1 in the supplementary info you can see Ust'-Ishim is closer to modern Africans in both haplogroup and PCA.

The big surprise for me here is the high heterozygosity level - although Stuttgart also shows higher heterozygosity than modern Eurasians I'd assumed this was due to the onset of farming and subsequent population growth. This result would indicate that modern humans in Ust'-Ishim's time were incredibly widespread and successful - not just a small isolated band of hunter-gatherers ekeing out a living.

Tobus said...

@agiering: Wouldn't this make Ust'-ishim similar to Amerindians, who also are more similar to East Asians, WHG, and Mal'ta (i.e., "ANE") than they are to Stuttgart (i.e., "EEF")?

Not really: UI is *equally* similar to East Asians, WHG, ANE and Amerindians and only slightly less similar to EEF. On the other hand Amerindians are *very* similar to East Asians and only a bit similar to ANE, with no similarity to WHG or EEF at all (note that Amerindian affinity with mixed WHG/ANE popluations like LB and Aj form a cline in line with the degree of ANE ancestry).

So UI isn't particularly like Amerindians at all, it's just the way you phrased it.

Hector said...

It seems that Ust-ishim man may belong to K2a(revised and extended version).
He does not share too many SNPs though probably meaning that his line branched off very early.

Magoon reportedly announced

CTS11667+
Z4842/M2308+
Z12216-
Z4845/M2313-
Z12176-
Z4952/M2339-
F650/M2346-

I once argued that NO line might have taken a northern route before the rest entered SE Asia.
TerryT was mighty upset for no reason when I said this but maybe I was right.

But still this is quite surprising. I thought of branching off somewhere in East India, not Afghanistan/Pakistan area as this new find may imply.

But still it is way too early to tell anything. Maybe Ust-Ishim man entered Siberia from the east and NO line may have taken the southern route as well.(counter-clockwise)

Hector said...
This comment has been removed by the author.
German Dziebel said...

@Tobus

You haven't studied http://anthropogenesis.kinshipstudies.org/2014/10/ancient-ust-ishim-dna-as-seen-from-the-americas/.

"On the other hand Amerindians are *very* similar to East Asians and only a bit similar to ANE, with no similarity to WHG or EEF at all.."

Amerindians are closer to modern Europeans than East Asians or Papuans. Modern Europeans have "Basal Eurasian" that UI does not have. A natural conclusion is that Amerindians straddle the ENA, ANE and BEu clusters. In good alignment with the mtDNA data.

"If you check figures 8.1, 9.1 and 10.1 in the supplementary info you can see Ust'-Ishim is closer to modern Africans in both haplogroup and PCA."

Closer than what? And what's closer to what: Africans to non-Africans maybe?

"So UI isn't particularly like Amerindians at all."

On the PCA from SI S10.1A and B (eigenvector 1) UI is the closest to SouthCentral Asians and Amerindians. Amerindians are also the closest to UI in terms of the degree of the preservation of the "Neandertal" chunks.

"The big surprise for me here is the high heterozygosity level."

An admixed population with ANE/WE and ENA ingredients, hence it's as close to MA-1/LB as it is to Onge. It's within the non-African heterozygosity spectrum, only as admixed as some West Eurasian groups.

terryt said...

"But don't forget that autosomal heritage is very different from uniparental haplogroups".

Yes. Many here and elsewhere seem to be forgetting that which is why they cannot see that Y-DNA K2 almost certainly originated in SE Asia. That is the reason Ust Ishim has minimal SE Asian a-DNA.

"This result would indicate that modern humans in Ust'-Ishim's time were incredibly widespread and successful - not just a small isolated band of hunter-gatherers ekeing out a living".

Likely to be the case. It would have been the later severe climate deterioration that forced much of the human population back south.

"TerryT was mighty upset for no reason when I said this but maybe I was right".

I doubt I was 'upset'. Surprised at the suggestion more like.

"But still it is way too early to tell anything. Maybe Ust-Ishim man entered Siberia from the east and NO line may have taken the southern route as well.(counter-clockwise)"

I would think so.

"Yet another reason for doubting that K2 and P migrating to northern and central Eurasia from southeast Asia".

I can't see how you can say that with any conviction. All K2 except K2b2 is SE Asian.

eurologist said...

Terry,

You misunderstood me - I was talking about the "Basal European"/ EEF component" - which is autosomally removed from all other Eurasians, extant or ancient as far as we know by now, and which y-DNA it could possibly be associated with.

Also, I agree with others that more work needs to be done to see how many upstream mutations Ust’-Ishim is away from basal P.

terryt said...

"It seems that Ust-ishim man may belong to K2a(revised and extended version).
He does not share too many SNPs though probably meaning that his line branched off very early".

Of course Ust'-Ishim's Y-DNA doesn't originate in SE Asia because K2 presumably first left South Asia via Southwest China, specifically Yunnan. But what the Ust'-Ishim Y-DNA does suggest is that K2a, otherwise known as NO, may also have moved considerably north of the Yunnan/Burma border region before N, O1, O2 and O3 formed. And that northward movement very likely happened after the ancestor(s) of K*, K2b, K2c and K2d had reached Sundaland and diversified there. In other words K2-M526 had parted company with K1-P326 it rapidly spread north and south all the way from Bali to Northern China.

Tobus said...

@German:
Amerindians are closer to modern Europeans than East Asians or Papuans

This kind of statement is inherently ambiguous - in future can you please make it clear if you mean "closer to YYY than XXX are" or "closer to YYY than to XXX". I'll assume you mean "closer than East Asians or Papuans are" in this case since the alternative is patently untrue.

A natural conclusion is that Amerindians straddle the ENA, ANE and BEu clusters

Amerindians don't have BEu.


Closer than what?

Than East Asians of course! Read the quote at the top of my previous post, that's what I'm responding to (doh!).

On the PCA from SI S10.1A and B (eigenvector 1) UI is the closest to SouthCentral Asians and Amerindians.

Eigenvector 1 is the X-axis (the one across the bottom). In S10.1A Central/South Asians, Europeans, Middle Easteners, North Asians, Oceanians and North Africans all score closer to UI than Amerindians do on the X-axis. On S10.1B (apart from a few outliers) Amerindians are a similar distance from UI as Europeans are. On both plots Central/South Asians are the closest (in some cases identical) to UI on that axis (and on the Y axis too!), so shouldn't you be positing that UI is similar to South Asians, not to Amerindians, based on this PCA?

It's within the non-African heterozygosity spectrum

It's actually just outside the modern non-African heterozygosity spectrum which is suprising to me (and no doubt to you since you have previous stated that low heterozygosity is an "ancestral condition").

As explained to you previously, if UI's increased heterozyogity is the result of admixture then we'd need an archaic population with an absolutely HUGE heterozygosity (resultant heterozygosity is the average of the two contributing populations, not the sum!).

Moreover, like Tianyuan, UI's position roughly central to all modern Eurasian populations indicates he is of a lineage ancestral to all of them - if he is from a different lineage to modern Amerindians (ie one branch of a common American ancestor) then we'd expect to see a closer affinity to Western Eurasians and the other descendants of the "Out of America" branch... particularly so if there was admixture with an extant archaic Eurasian population as you suggest.

German Dziebel said...

@Tobus

"This kind of statement is inherently ambiguous - in future can you please make it clear if you mean "closer to YYY than XXX are" or "closer to YYY than to XXX". I'll assume you mean "closer than East Asians or Papuans are" in this case since the alternative is patently untrue."

Good attempt at filibustering! My statement was not ambiguous and it's absolutely true. Raghavan 2013, Fig. 3b: "all of the 17 tested western Eurasian populations are closer to Karitiana than to Han Chinese."

"Amerindians don't have BEu."

Sure thing. This is because BEu is derived from Amerindians, hence it's post-Amerindian and not proto-Amerindian.

"On both plots Central/South Asians are the closest (in some cases identical) to UI on that axis (and on the Y axis too!), so shouldn't you be positing that UI is similar to South Asians, not to Amerindians, based on this PCA?"

Yes, South/central Asians are closer to UI than Amerindians. This is only natural considering the much greater geographic proximity between UI and South and central Asia compared to the New World. But out of all the geographically most distant modern populations Amerindians are the closest to UI.

"It's actually just outside the modern non-African heterozygosity spectrum."

MLRO is a more reliable measure than GATK and by MLRO measures French A and Sardinian A have higher heterozygosity values than UI. For comparison, GATK heterozygosity in Loschbour is higher than GATK heterozygosity in Karitiana. Considering that you used MLRHO values to argue that Loschbour "beats" Karitiana on homozygosity, you should probably stick with MLRO for UI as well. Just to avoid double standards.

"have previous stated that low heterozygosity is an "ancestral condition".

Yes it must be considering that Denisova and Neandertals are very homozygous. This helps explaining UI heterozygosity as product of admixture, which in turn explains its equal distance to both MA-1/LB and to East Eurasians.

"As explained to you previously, if UI's increased heterozyogity is the result of admixture then we'd need an archaic population with an absolutely HUGE heterozygosity (resultant heterozygosity is the average of the two contributing populations, not the sum!)."

heterozygosity in admixed populations is always higher than in non-admixed ones. Population subdivision measured by Fst, etc. is the reverse of that. So two populations with "HUGE" Fst values would generate a population with UI levels of heterozygosity. Populations with high Fst are Denisovans, Neandertals, Amerindians and then decreasingly all the way down to Africans.

"UI's position roughly central to all modern Eurasian populations indicates he is of a lineage ancestral to all of them - if he is from a different lineage to modern Amerindians (ie one branch of a common American ancestor) then we'd expect to see a closer affinity to Western Eurasians and the other descendants of the "Out of America" branch... particularly so if there was admixture with an extant archaic Eurasian population as you suggest."

A central position is equally compatible with admixture. And why do you require a closer affinity to West Eurasians? In its non-Neandertal-related loci, UI is apparently closer to East Eurasians (Han, Onge) than to Amerindians (see Fu et al. 2014, Fig. 3). But Amerindians, as discussed, are closer to MA-1, LB and modern Europeans than East Eurasians. This means that UI sits on one branch of an ancient Amerindian-derived population but not on the other. But since it's as old as 45,000 years, it's pulled toward the ancient representatives of the West Eurasian branch as well but not as much as Amerindians are.

Gregory76 said...

Terry,

I maintain that K2 diversified in and spread from Central or Southwest Asia for the same reasons I have been maintaining it, including:
1. this theory requires less migration than the theory that has it diversify in and spread from Southeast Asia in an era when staying alive for a society was very hard, and staying live will migrating still harder.
2. O is bearer of the Mongoloid phenotype from Mongolia to Indonesia and out to Polynesia, and the Mongoloid phenotype developed in response to a very cold climate (two points which, as I remember, you agree with me on), and that was most likely Siberia (whereas you say it was Tibet).

Yes, several of branches of the macro-haplogroup that began with K2, including the oldest ones, are found in Southeast Asia, and both of those points do count strongly in favor of the Southeast Asian theory, but the location of a majority of subgroups of group and of the oldest subgroup is not always the ancestral home, and their weighy reasons on the other side.

As to the claim that five of K2's branches developing in Southeast, that claim, too, remains to be proved: I do not concede that K2a (NOX) and K2b(PMS) developed there.
That NOX developed there is especially doubtful: the great majority of people with N live in North Asia or Europe, and I think we both believe A originated in Eastern, not Southeastern, Asia.
As to PMS, M and S originated in Southeast Asia, but it seems about as doubtful that P did as it is that NOX did: the great majority of people with Q (in the Old World, where it originated) are found in North Asia, and while R is found in India and Iranian, most people with R live or have lived in Europe, west central Asia and western Siberia. So while a majority of the lineages directly descended from PMS originated in Southeast Asia, it is only a majority of 2 out of 3. So it is easy to imagine PMS giving rise to P, M and S in Central Asia and then a branch of PMS proceeding with M and S to Southeast Asia, or else giving rise to P in Central Asia and then proceeding to Southeast Asia and giving rise to M and S there.

Similarly, it is easy to imagine K2 giving rise to K2a, K2b, K2c and K2d, and Ust-Ishim, in Central Asia and then a branch of K proceeding with K2c and K2d to Southeast Asia, or else giving rise to K2a, K2b and Ust-Ishm in Central Asia and then proceeding to Southeast Asia and giving rise to K2c and K2d there.

Also, it was you who pointed out the short time the ancestors of Ust-Ishim had to migrate from Southeast Asia to Central Asia.

Finally, you said

“...Ust Ishim has minimal SE Asian a-DNA” by way of supporting you theory, but that sounds like a reason for thinking that K2 did not originate in SE Asia, since Ust-Ishim, a very early example of K2, did not have SE Asian a-DNA.

gnomon said...

@capra internetensis @Tobus

Thank you for your reply and your understanding of my comment. Our difference is the length of separation time between Africans and non-Africans. From the latest Y chr estimate by Ponzik et al 2013, the time to MRCA is 140000 years. Both of you have used time similar to that. But that time is for the split of A and BT hyplotypes, with B almost only found in Africans. The separation time one should use however is for the split of F (common in nonAfr but much less so in Afr) and E (representing the majority of Afr or Bantu and much less common in Non-Afr), which is in fact closer to the 50000 years used in my original comment.Another point to keep in mind is that Non-Africans did not exit Africa until 50000 years ago according to the latest model.When everyone was in Africa, it is hard to imagine any barrier to prevent hybridization. So, the more correct estimate for the separation time between African and Non-African should be closer to 50000 years rather than >100000 years.

Tobus said...

@gnomon: So, the more correct estimate for the separation time between African and Non-African should be closer to 50000 years rather than >100000 years.

I think you missed the main point that you need to count the time on both branches, not just the non-African one. Even using a divergence as recently as 50kya ago there'd still be 100ky of difference between modern African and modern non-Africans - 50ky from modern non-Africans back to the MCRA and then 50ky from the MCRA to modern Africans. On this time scale there'd be 55ky difference between modern Africans and UI (5ky back to the MCRA and 50ky to modern Africans).

@German:
My statement was not ambiguous...

Rhagavan explicitly states "than to the Han Chinese", whereas your statement leaves that "to" out and so could potentially be parsed as "than the Han Chinese are". If you want to be unambiguous, explicitly state either "closer to X than to Y" or "closer to X than Y are", depending on what you mean.

This is because BEu is derived from Amerindians, hence it's post-Amerindian and not proto-Amerindian.

Um... I see... "straddle the cluster" wasn't a particularly apt choice of wording then, was it.

Yes it must be [ancestral] considering that Denisova and Neandertals are very homozygous

Denisovans and Neandertals aren't ancestral to modern humans.

MLRO is a more reliable measure than GATK

By either measure UI has an exceptionally high heterozygosity compared to other ancient HG populations, I would have expected levels similar, or even lower than, his WHG ancestors like Loschbour. Instead we have levels up with the Neolithic farmers and their descendants. Quite unexpected.

heterozygosity in admixed populations is always higher than in non-admixed ones

... and yet the EEF/WHG admixed French have a heterozygosity in between EEF and WHG. Do you have any real world data that shows an admixed population having a higher heterozygosity than each of it's contributors? No? I didn't think so.

And why do you require a closer affinity to West Eurasians?

For the same reason in my explanaton to gnomon above - you have to count both branch lengths. If, say, a hypothetical population in America splits at 50kya and one group goes West into Eurasia while one group stays in America, then all the descendants of that Western branch will be closer to each other than any of them are to the branch that stayed. From UI's point of view, branch length to modern Eurasians will be 45ky, while branch length to modern Amerindians will be 55kya (5kya back to the common ancestor plus 50k to modern Amerindians). On top of this you are suggesting heavy admixture in the Western branch making it even *more* divergent from the unadmixed American branch. The fact that UI is essentially central to all these non-Africans indicates he existed before any of their individual lineages had branched off - that's the only way he can have equal branch length to all of them.

UI is apparently closer to East Eurasians (Han, Onge) than to Amerindians (see Fu et al. 2014, Fig. 3)

See Table S11.2 in the SI - the D(Onge/Han, Karitiana, UI, Chimp) results show no significant difference (Z scores less than 1!).. and read point 2 of the Results underneath it: "Specifically, we find that none of these samples shares significantly more alleles with Ust’-Ishim than any other"

gnomon said...

@Tobus
If you agree with me that Afr and non-Afr separated 50000 years ago, then the distance between UI and Afr is (50000 + 5000) x rate, and the distance between East Asians and Afr is (50000 + 50000) x rate. Hence UI should be nearly 2 fold less distant to Afr than East Asians are, as stated in my original comment.

Along the same line, another surprise I would like to bring up is the fact that UI is an outlier to living East Asians on a PCA plot as shown by the paper. But UI should not be an outlier! The distance between UI and East Asians should be almost 2 fold less distant than between certain pair of East Asians who could have a distance of (50000 + 50000) x rate, whereas the distance between UI and East Asians is (50000 + 5000) x rate.

terryt said...

"that sounds like a reason for thinking that K2 did not originate in SE Asia, since Ust-Ishim, a very early example of K2, did not have SE Asian a-DNA".

Don't forget that Y-DNA is, in effect, just one of the tens of thousands of human genes. Haploid DNA can easily introgress into pre-existing populations without changing the a-DNA significantly.

"this theory requires less migration than the theory that has it diversify in and spread from Southeast Asia in an era when staying alive for a society was very hard, and staying live will migrating still harder".

The recovery of ancient genetic samples shows humans were much more mobile than anyone had suspected. I think it is far more likely than ancestral forms of particular haplogroups moved and localised variations then formed, rather than that populations moved with fully-formed pre-existing diversity. Ancient humans were much more mobile than most give them credit for.

"O is bearer of the Mongoloid phenotype from Mongolia to Indonesia and out to Polynesia, and the Mongoloid phenotype developed in response to a very cold climate (two points which, as I remember, you agree with me on), and that was most likely Siberia (whereas you say it was Tibet)".

I do agree with you on that. I think the EDAR370A mutation is good proxy for the development and expansion of the Mongoloid phenotype. The authors of the original paper place its origin at the Eastern Tibet/China border region. That explains:

"That NOX developed there is especially doubtful: the great majority of people with N live in North Asia or Europe, and I think we both believe A originated in Eastern, not Southeastern, Asia".

NOX (or K2a) must have moved up the China/Tibet border region from the Burma/China region. At the northern end it picked up the Mongoloid phenotype and began a new expansion.

"Yes, several of branches of the macro-haplogroup that began with K2, including the oldest ones, are found in Southeast Asia, and both of those points do count strongly in favor of the Southeast Asian theory"

Yes.

"As to the claim that five of K2's branches developing in Southeast, that claim, too, remains to be proved: I do not concede that K2a (NOX) and K2b(PMS) developed there".

As I pointed out technically K2a did not develop in SE ASia but it is very difficult to conclude other than that K2*, K2b, K2c and K2d developed anywhere other than in SE Asia. Which means K2b2 must have split from K2b1 in that region.

"As to PMS, M and S originated in Southeast Asia, but it seems about as doubtful that P did"

You can't just consider M and S here although I agree K2b2 is on a separate line from K2b1. We have four K2b1 lines, all confined to SE Asia: M, K2b1b, K2b1c and K2b1b (which includes S). That makes it difficult to place K2b2 anywhere else with any confidence at all.

"the great majority of people with Q (in the Old World, where it originated) are found in North Asia, and while R is found in India and Iranian, most people with R live or have lived in Europe, west central Asia and western Siberia".

Yes. All downstream from K2b2. I'm sure that any population capable of crossing Wallace's Line in a boat would have been quite capable of moving up the Ganges with ease. And rapidly as the nearby mainland was already occupied. Which explains:

"Also, it was you who pointed out the short time the ancestors of Ust-Ishim had to migrate from Southeast Asia to Central Asia".

Simply explained.

"So it is easy to imagine PMS giving rise to P, M and S in Central Asia and then a branch of PMS proceeding with M and S to Southeast Asia"

I find that very difficult to imagine. Are you suggesting K2b1b, K2b1c, M as well as four branches of K2b1a had formed long before they reached Wallace's Line? Leaving no stragglers along the way? Remember we have stragglers of R and even Q in South Asia.

eurologist said...

The separation time one should use however is for the split of F (common in nonAfr but much less so in Afr) and E (representing the majority of Afr or Bantu and much less common in Non-Afr), which is in fact closer to the 50000 years used in my original comment. Another point to keep in mind is that Non-Africans did not exit Africa until 50000 years ago according to the latest model.

Gnomon,

What "latest model?" The best current dates for ooA are ~125,000 - 105,000 ya. Also, to date there is no evidence at all that E originated in Africa. Conversely, there is a ton of evidence that most Africans have West Eurasian autosomal contributions that pre-date agriculture and y-DNA R and J impacts.

y-DNA haplogroup E may have arisen at the NE African/ SW Asian interface, if not solely in SW Asia - but it definitely is way, way older than 50,000 years.

Slumbery said...

Tobus

"As explained to you previously, if UI's increased heterozyogity is the result of admixture then we'd need an archaic population with an absolutely HUGE heterozygosity (resultant heterozygosity is the average of the two contributing populations, not the sum!)."

Not to join up your discourse with GB, but why would it be then average? This is counter-intuitive for me. Would the fusion of two completely homozygous populations not create heterozygosity as far as the two starting populations have a sufficient genetic difference? In case of the fusion of two strongly divergent populations the starting heterozygosity of each population certainly have an effect but it is possibly not even the main factor of the resulting heterozygosity (it may be, but not necessarily). Even if there was significant loss of alleles during an unequal fusion, the resulting heterozygosity could be even higher than the sum as far as the difference between the two populations is bigger than the heterozygosity of each.
(Again, not necessarily so and we know that probably most of the Neandertal alleles were lost, but this is still a possibility.)

Hamar Fox said...

Hector,

The author may have made an error in not taking it into the consideration that Malta boy etc. are supposed to be closer to this man than modern populations are to him, all other things being equal. This would mean that Ust Ishim man may be closer to East Eurasians already.

A poorly expressed point, and I understand you're basically just sulking that the results didn't go your way, but it seems here that you're scraping the barrel along the lines of, 'If ancient East Eurasians were compared with ancient West Eurasians, the affinity would possibly/probably/hopefully be with East Eurasians'.

But shared drift is shared drift. Unless modern East Eurasians have admixture with a population that has radically less shared drift with UI, such as modern, basally admixed West Eurasians or sub-Saharans (which they don't, and you wouldn't want to admit it if it were true anyway, at least not in the former case), or they've undergone intense drift that masks shared drift with other populations (which we can probably discount considering the diversity of ENA populations considered in this study and the uniformity of the findings for all of them), then I don't see how the 'real situation' could be significantly different from this paper's conclusions. If I understood it correctly, your point could apply to PCA results at best.

This is the most charitable interpretation of your comment, mind, since at first it looked as though you were arguing that UI was East Eurasian because it was autosomally most similar to ancient West Eurasians (which doesn't seem to be a correct point anyway) -- and as used as I am to you being wrong about things, a logical error of that magnitude would just be too much for me to believe.

But these findings are basically consistent with what I predicted in our last conversation: ancient populations whose haplotypes (or descended haplogroups) are entrenched in populations at opposite extremes of variation are likely to have been autosomally neutral (or close to neutral) between those modern extremes, unless some kind of shift eastward or westward can be shown to correlate with these haplogroups relative to haplogroups that have no such ambiguous origin (for instance, if R1b-dominance in a population could be shown to correlate with an eastward shift relative to, say, I-dominated populations -- which, obviously, Basques show isn't the case).

Rokus said...

'Admixture will result in a heterozygosity higher than the less heterozygous contributor, but *lower* than the more heterozygous one'

No, the offspring of heterozygous twins - if brothers or sisters could mate - would probably be close to homozygous, while the offspring of different homozygous species - if they could mate - would result in the highest possible level of heterozygosity.

'As explained to you previously, if UI's increased heterozyogity is the result of admixture then we'd need an archaic population with an absolutely HUGE heterozygosity (resultant heterozygosity is the average of the two contributing populations, not the sum!).'

All the contrary, UI's increased heterozygosity is most probably indicative of the primordial hybridization I always imagined to be the source of modernity: 'Expanding Hybrids And The Rise Of Our Genetic Common Denominator'

Rokus said...

'All K2 except K2b2 is SE Asian.'
Only nowadays. However, just remember that modern populations must have acquired their Y-DNA at some point of time, or several times in consecutive waves. Current distributions won't tell you from where. A much better reference is UI, being close to ancestral to all K2, and still far from being SE Asian.

ryukendo kendow said...

@ Gregory76
I think this discovery, if anything, weakens the basis for any discussion about human migrations based on haploid genetics, and strengthens the case for discussion based on autosomal genetics.

The SNP table shows that K14 did not belong to the same C as La-Brana, but rather a basal version of C1b, which is not present after him in Europe and is found in India, Oceania and East Asia if I'm not wrong.

So we have Ust-Ishim on the ENA line with a Yhap that roots him in East Asia, and another guy with WHG with a Yhap that roots him further east as well. And further south, because on his way here he grazed Basal populations.

Last of all we have Kostenki's skull, and it is not caucasoid in any sense, while the autosome is. East Asians may have branched into a separate population with its own Hg pool prior to any physical adaptation.

Hamar Fox said...

Because I like revelling in Hector's misfortune, I should quote what I said in a 'conversation' with him in June:

"Alternatively, you may have a West-East [autosomal] division developing fairly early in Eurasian history, and perhaps these populations 'mopped up' 'neutral' lineages whose descendants ultimately fell wholly within the range of West Eurasian diversity (R1, maybe R completely, and Q) or wholly within the range of East Eurasian diversity (N, O)."

This is, of course, precisely what we [I]do[/I] see with Ush-Ishtim, and the fact that within eight thousand years of UI there were unambiguously West Eurasian populations (Kostenki) and unambiguously East Eurasian populations (Tianyuan) suggests that it's entirely possible that autosomally neutral UI-like populations shared Eurasia with diverging East and West Eurasian populations 45,000 years ago and that lineages to which K is ancestral correlate only with a muted shift east or west (in West and East Eurasians respectively), if any shift can be detected at all, because of the autosomal neutrality of the source population.

Still, even though I predicted it, I'm still surprised just [I]how[/I] neutral UI turned out to be.

'Australoid' populations likely have the greatest share of 'neutral Eurasian' ancestry in modern populations, which is likely why in PCAs they are much closer to the centerpoint of E. and W. Eurasians than East Asians are. Denisovan admixture possibly obscures this in Fst estimates. Some of Eurogenes' TreeMix runs show an admixture edge slightly upstream from modern Papuans to UI, presumably to separate the 'neutral' ancestry in Papuans from the 'East Eurasian' element.

terryt said...

@ Rokus:

"Only nowadays [All K2 except K2b2 is SE Asian]".

To me it seems that haplogroup extinction has seldom occurred through expansion of new haplogroups, except over great time lengths. We find remnant haplogroups in many regions. Where we find obvious haplogroup extinction is in regions where environmental change has been significant and the extinction is probably the result of overall human extinction.

"A much better reference is UI, being close to ancestral to all K2, and still far from being SE Asian".

Evidently not 'close to ancestral to all K2'. Close to ancestral K2a only, which I agree is not technically SE Asian. But fairly close to it.

"UI's increased heterozygosity is most probably indicative of the primordial hybridization I always imagined to be the source of modernity"

Agreed. In fact I think the role of hybrid vigour in human expansions (and in that of many other species) has been consistently overlooked.

@ ryukendo kendow:

"I think this discovery, if anything, weakens the basis for any discussion about human migrations based on haploid genetics, and strengthens the case for discussion based on autosomal genetics".

I disagree. What it shows, in my opinion, is that haploid expansion usually largely overlays pre-existing autosomal genetics. Elements of previous populations survive even though the proportion of their haploid genetics can be greatly reduced.

"K14 did not belong to the same C as La-Brana, but rather a basal version of C1b, which is not present after him in Europe and is found in India, Oceania and East Asia if I'm not wrong".

That seems to be correct.

@ Hamar Fox:

"'Australoid' populations likely have the greatest share of 'neutral Eurasian' ancestry in modern populations, which is likely why in PCAs they are much closer to the centerpoint of E. and W. Eurasians than East Asians are. Denisovan admixture possibly obscures this in Fst estimates".

As far as I'm aware Papuans have far more Denisova than do Australoids. It pays to keep in mind the two a re more different than most like to believe. I agree, however, with the rest of your commnet.

German Dziebel said...

@Tobus

"Denisovans and Neandertals aren't ancestral to modern humans."

We don't know that. Those are the only two hominin populations that we know we share genetic ancestry with. It's an open question whether we are related to them through common descent, admixture or both. The fact that Amerindian population structure is the closest to their population structure, among all modern human populations, plus Amerindians share ancestry with them on the order that's higher than modern West Eurasians or Africans plus Amerindians has the worldwide highest levels of linguistic diversity (and language is the chief sign of behavioral modernity) seems to be a consistent signal supporting the idea that modern humans descended from an East Eurasian hominin species related to Denisovans and Neandertals and were isolated in the Americas before re-expanding to the Old World as modern humans and re-admixing with Old World populations such as Denisovans, Neandertals and African archaics.

"See Table S11.2 in the SI - the D(Onge/Han, Karitiana, UI, Chimp) results show no significant difference (Z scores less than 1!).. and read point 2 of the Results underneath it: "Specifically, we find that none of these samples shares significantly more alleles with Ust’-Ishim than any other"

Within the theoretical framework that the authors have in mind those differences are indeed negligible. But when out-of-America is taken into consideration then the stronger pull of Amerindians (compared with East Asians) toward West Eurasians vs. Ust-Ishim and of East Asians such as Onge and Han (compared to Amerindians and West Eurasians) toward Ust-Ishim (Fig. 3) becomes meaningful. Correspondingly, UI is Y-DNA hg X that leads to East Asian NO bit not to Amerindian-West Eurasian P. All of this is consistent with one of the predictions of out-of-America whereby Amerindians split off prior to the West Eurasian vs. East Asian split. You can't explain this as admixture between very old East Asians (Tianyuan) and very old West Eurasians ("Basal Eurasians" or Kostenki) because a) Amerindians are too homozygous; b) they occupy their corner in PCA plots but there were no New World hominins that they could have admixed with and that would have pulled them apart from Old World populations, so it must be their "authentically modern human" divergence and not the "secondary divergence" on "archaic steroids" seen in Papuans or Africans; c) geographically it's unlikely that such dispersed groups would uniquely come together to enter America at a very early date. So out-of-America provides a very solid alternative to mainstream meanderings.

"By either measure UI has an exceptionally high heterozygosity compared to other ancient HG populations, I would have expected levels similar, or even lower than, his WHG ancestors like Loschbour. Instead we have levels up with the Neolithic farmers and their descendants. Quite unexpected."

It's an admixed population.

" The fact that UI is essentially central to all these non-Africans indicates he existed before any of their individual lineages had branched off - that's the only way he can have equal branch length to all of them."

No, it's an admixed population that's why it's in the middle. It's admixed between ancient East Asians such as Tianyuan and ancient West Eurasians such as MA-1 or LB (without the Amerindian-BEu/Middle East-Stuttgart-French lineage) and possibly Neandertals. Its levels of heterozygosity suggest this as well. MA-1 is also consistently in-between West Eurasians and Amerindians just like Aleutians who have a large proportion of offspring from post-1492 marriages between Native Americans and Europeans.

terryt said...

A further comment in relation to my reply to Rokus's comment, 'Only nowadays'. Of course minority haplogroups get drifted out over time. Those minority haplogroups can be the result of either incoming haplogroups outnumbering indigenous ones or the incoming group being small enough to be outnumbered by indigenous haplogroups. But such drift can be a very long process.

Tobus said...

@Slumberry: Would the fusion of two completely homozygous populations not create heterozygosity as far as the two starting populations have a sufficient genetic difference?

Theoretically yes, but we're talking about reality here - no such populations could ever exist. In real world examples there are in fact very few completely homozygous derived alleles - nearly all surviving mutations are non-functional (so not subject to selection) and of the small percentage that are, even the most highly selected ones often only approach fixation, with frequencies in the high 80s or 90s. So the vast bulk of derived sites in any population are heterozygous to some degree - some of them will be as high as the 90-99% and virtually ensure a new heterozygous site, but an equal number are as low as 1-10%, virtually ensuring a heterozygous site disappears. I think you understand that the chance of a derived allele at a heterozygous site passing to a child is 50%, which means that the population-wide frequency of the derived allele at heterozygous sites gets halved by admixture (assuming 50/50 contribution for simplicity). When we sample a random individual a few hundred generations down the track the result will be roughly the average of the two contributing populations (in proportion to their contribution).

Intuitively, consider that admixture homogenises differences between populations. It results in a range of phenotype that is always between the contributing populations, never more extreme than either - a mix between a large-nosed population and a small-nosed population will never result in a child with a nose that is bigger (or smaller) than either parent. Because it only takes half of each parent, admixture averages out the differences between populations, it doesn't stack them on top of each other. A heterozygosity somewhere in between that of your parents is completely expected.

Mathematically, consider a jar full of blue marbles but with every 20th one red. If we take half this out and replace it from another jar of blue marbles that has every 10th one red, the result will be a jar with roughly every 15th marble red - not 30.

And last but not least... empirically, I have been unable to find any real world example of an admixed population having higher heterozygosity than both it's parents. If you are able to show me one, I'll change my mind.

Tobus said...

@Rokus:
No, the offspring of heterozygous twins - if brothers or sisters could mate - would probably be close to homozygous...

Probably not. Assuming you mean identical twins with the same heterozygous sites, the chance of any heterozygous site being homozygous in their offspring is 50% (25% AA, 25% BB), so we'd expect the offspring to have roughly half the number of heterozygous sites as their parents. Given there are millions of heterozygous sites in modern humans, the probability of them having a child with almost no heterozygous sites is approaching zero.

while the offspring of different homozygous species - if they could mate - would result in the highest possible level of heterozygosity.

Yes, *if they could mate*. We're talking about populations which *can* mate (and very successfully!), which precludes any large difference in homozygous sites.

All the contrary, UI's increased heterozygosity is most probably indicative of the primordial hybridization I always imagined to be the source of modernity

Probably not - I've explained above why admixture won't increase overall population heterozygosity. Is this just some assumption you've made or can you show me a real world example of a hybridisation that resulted in heterozygosity level higher than both the source populations?

Gregory76 said...

K ex LT Update

Terry,

I see that I underestimated the strength of my own theory in my last post.
I conceded that 2 of 3 descendents of MPS (K2b)--namely, M and S—were from Southeast Asia/New Guinea.
However, I see now that M and S are both branches of K2b1. So only 1 of the 2 descendents of K2b are mainly from Southeast Asia/New Guinea.
So that leaves us with these conclusions about the branches of K2:

K2c—Southeast Asia
K2d—Southeast Asia
Ust Ishim—North/Central Asia
K2a (NOX)—mostly northern, with its Southeast Asian branch arriving there from East Asia
K2b (MPS)—two sub-branches:
K2b1—Southeast Asia/New Guinea
K2b2 (P)—mostly northern or western distribution

So that is two branches in Southeast Asia, two mostly northern or western, and one with a sub-branch in each area. So we can no longer say that a clear majority of the branches of K2 originated in Southeast Asia.

K2c and K2d may have arisen from K2 in Southwest Asia, but it could also be the case that K2C and K2d arose from K2 after K2 had expanded to Southeast Asia.

As for stragglers, it is very likely that large number of small bands of people became extinct while migrating, since it was easy for a small population to be wiped out by illness, war, natural disaster, etc.

Rokus said...

'Of course minority haplogroups get drifted out over time. Those minority haplogroups can be the result of either incoming haplogroups outnumbering indigenous ones or the incoming group being small enough to be outnumbered by indigenous haplogroups. But such drift can be a very long process.'

Indeed we are talking about processes that lasted thousands of years. Just remember that mathematically, with a growth factor of 50% per generation, ie. on average three children by female (eg. less than Africa nowadays! http://data.worldbank.org/indicator/SP.DYN.TFRT.IN), it takes just one thousand years for a closed population to grow from one couple to three million descendants - more than enough to replace all humans in prehistory. Now, that would imply drift and extinction! Actually, there is not any reason to assume something like a numeric 'balance' between haplogroups, or any other sort of quotum to protect prehistoric minorities from extinction - except through the mechanism of hybridization. Neither may we assume the Y-chromosome exempt of any favorable mutation for tens of thousands of years. Natural selection is most likely to have had a more egalitarian effect on populations surviving in arctic or arid climates. In other words, we can safely discard mere statistics to explain current distributions of haplogroups. Moreover, neither complicated back-and-forth migrations nor simple one way SE Asia expansions towards a Siberian target between UI and Mal'ta, only involving hg K, would apply as parsimonious scenarios.

Rokus said...

'Given there are millions of heterozygous sites in modern humans, the probability of them having a child with almost no heterozygous sites is approaching zero.'
The heterogenous offspring of two homozygous individuals from very distinct pupulations could mate again and thus easily restore large tracts of homozygous DNA. Within three generations there would thus be a shift from very homogenous to very heterogenous to much less heterogenous.

'I've explained above why admixture won't increase overall population heterozygosity. Is this just some assumption you've made or can you show me a real world example of a hybridisation that resulted in heterozygosity level higher than both the source populations?'
Examples abound. Actually, heterozygosity is frequenty caused by admixture of distinct populations that more often than not are considered homozygous to some degree. Increased levels of heterozygosity can be measured within any mixed population, like the Caribbean, Mexico or Brazil.

terryt said...

"So that is two branches in Southeast Asia, two mostly northern or western, and one with a sub-branch in each area. So we can no longer say that a clear majority of the branches of K2 originated in Southeast Asia".

Not quite correct. Your own list shows three branches in SE Asia unless you're claiming New Guinea as being completely independent of SE Asia. In fact K2c and K2d are specifically 'island SE Asia', Bali and Java respectively. The same holds for K2* which may or may not be a single haplogroup, perhaps one each in Sumatra and Sulawesi. That is three or even four, specifically SE Asian branches. Turning to K2b. One branch, K2b1, is also specifically island SE Asia (and New Guinea). That makes three (or four) and a half SE Asian haplogroups. And not to forget that K2a has a very strong presence in SE Asia although that is probably the result of later southward movement. But NOX could hardly be considered 'northern'. Perhaps northern Chinese at most although we do have the closely related Ust'-Ishim branch definitely northern but unrelated to the more definitely northern K2b2(P).

"K2c and K2d may have arisen from K2 in Southwest Asia"

It is extremely doubtful they would have become sorted as to which island they ended up on if such was the case. I agree it is more likely that they differentiated from each other once they had arrived on said islands, but they would have arrived as generalised K2.

"As for stragglers, it is very likely that large number of small bands of people became extinct while migrating, since it was easy for a small population to be wiped out by illness, war, natural disaster, etc."

True to some extent but in that case we should expect to find the odd pocket of surviving stragglers along the route. And a population migrating through a previously unoccupied region is unlikely to have remained 'a small population' for very long.

German Dziebel said...

@Tobus

"Is this just some assumption you've made or can you show me a real world example of a hybridisation that resulted in heterozygosity level higher than both the source populations?"

There are plenty of studies showing correlation between heterozygosty and admixture across species. At some point there was a regression analysis done in Caucasian-admixed vs. unadmixed South American tribes and admixture accounted for some 80% of heterozygosity variance.

"Theoretically yes, but we're talking about reality here - no such populations could ever exist."

I just gave you one. And it's right on the surface.

"Intuitively, consider that admixture homogenises differences between populations."

Sure it does. It levels out intergroup or between-group differences (Fst, etc.) and increases intragroup, within-group differences. http://rosenberglab.stanford.edu/papers/BocaRosenberg2011-TPB.pdf. Africans are highly admixed and heterozygous, Amerindians are the least admixed and most homozygous among modern populations. Amerindians are too homozygous to allow for a mix between more heterozygous East Asians and West Eurasians.

"And last but not least... empirically, I have been unable to find any real world example of an admixed population having higher heterozygosity than both it's parents. If you are able to show me one, I'll change my mind."

The last thing I want is for you to change your mind. But it'd be good if you supported your speculations with an actual example of an admixed population with lower heterozygosity than both its parent populations.

German Dziebel said...

here's one of many papers that shows the role of outbreeding and mixing in driving heterozygosty. http://www.nature.com/ejhg/journal/v16/n9/full/ejhg200848a.html

Tobus said...

@German:
We don't know that

Yes we do - we have sapiens samples that date before, during and after the Neanderthal/Denisovan samples that show an affinity to modern humans tens if not hundreds of times greater than any of the N/D samples do. The UI sample is a much, Much, MUCH! better estimation of our ancestral state than the dying dregs of a completely different lineage.

Amerindian population structure is the closest to their population structure, among all modern human populations

We've been down this road before German: No modern human population has a population structure remotely similar to N/D. Amerindians are actually closer in population structure to modern Africans than they are to N/D and very, very similar (on that scale) to East Asians, Papuan, Australians etc. You are just (and still!) playing semantics.

plus Amerindians share ancestry with them on the order that's higher than modern West Eurasians or African

Not "higher" just "less recombined", consistent with Amerindians going through a population bottleneck that the others didn't.

plus Amerindians has the worldwide highest levels of linguistic diversity

We've been down this road before too German: Linguistics has a horizon only a few thousands years into the past. We're talking about events at least an order of magnitude further back than that. Linguistics can't help you understand the origins of home sapiens, only confuse you.

seems to be a consistent signal supporting the idea

The only thing "consistent signal" is that of being misguided, as demonstrated by the three points above.

and re-admixing with Old World populations such as Denisovans, Neandertals

.. which would bring Old World populations closer to N/D than New World ones are, and yet your whole thesis is based on the New World populations being closer! Did I hear someone say "consistency"?

Within the theoretical framework that the authors have in mind those differences are indeed negligible. But when out-of-America is taken into consideration... it becomes meaningful

No German, the Z-score is an empirical measure of statistical significance, it doesn't change with your varying "theoretical frameworks". You Arts majors sure do crack me up sometimes.

It's an admixed population.

Ahhh, "admixture"! The magical glue that holds all crackpot theories together! We've been down this road before German: Admixture doesn't increase overall population heterozygosity. As I asked you previously, if you think I'm wrong please provide a real world example of an admixed population that has a heterozygosity higher than both of it's contributors. There are lots of examples where they don't.

No, it's an admixed population that's why it's in the middle

More admixture! This stuff is amazing! Just the right amount of just the right components in just the right ratios to make it *look* just like he's ancestral to everybody, but really it's the other way round! Incredible!

Seriously though, if UI really is an admixture of two Eurasian branches, he'd be pulled closer to them and away from modern Amerindians... think about branch lengths. You need a scenario where UI's branch length to each modern (non-BEu) population is the same. Ancestral to all fits this bill and, given his age, is both logical and parsimonious.

Slumbery said...

Tobus

Thank you for the detailed answer. I am still not convinced.

"Mathematically, consider a jar full of blue marbles but with every 20th one red. If we take half this out and replace it from another jar of blue marbles that has every 10th one red, the result will be a jar with roughly every 15th marble red - not 30.

And last but not least... empirically, I have been unable to find any real world example of an admixed population having higher heterozygosity than both it's parents. If you are able to show me one, I'll change my mind."


You choose a mathematical example that completely the opposite of my starting point. Let's say we have only fixed alleles in both populations, but different alleles fixed in the two. We mix a jar of 90 blue balls with a jar of 20 red balls. We got a heterozygous population from two homozygous one, at least for one particular gene.

We can define this as "inter-population heterozygousity" between two populations that otherwise have low intra-population heterozygousity.

Now, in overall (for all genes) this does not happen in modern humans as you pointed out. Well, I am pretty sure I could dig out cases where the mixed population is more heterozygous that the parent populations _for one or a few particular gene_, but you still have a point here about the overall heterozygousity.

The(present day and recent) modern humans are overall very homogeneous with relatively recent strong gene flows. This means that between any two populations of modern humans overall inter-population heterozygousity is always lower that the intra-population ones. Typically much lower.

If we conclude (as you do) that this is the only realistic starting position then of course everything work as you said. So it comes down to one question: Are you sure that this is also true in the modern humans vs. Neanderthals or modern humans vs. Denisovans relation?

terryt said...

"Just remember that mathematically, with a growth factor of 50% per generation, ie. on average three children by female (eg. less than Africa nowadays! http://data.worldbank.org/indicator/SP.DYN.TFRT.IN), it takes just one thousand years for a closed population to grow from one couple to three million descendants"

It is extremely unlikely human populations have ever grown that rapidly. Even in the case of New Zealand where people were able to move into a previously uninhabited region the population didn't grow near that rapidly. Resources rapidly give out unless continuous expansion is possible.

"Now, that would imply drift and extinction!"

Usually random as far as haplogroups are concerned. In fact, as I said, minority haplogroups would get drifted out first.

"Neither may we assume the Y-chromosome exempt of any favorable mutation for tens of thousands of years".

'Selection' on haplogroups is very unlikely to be genetic. Technological of cultural ability associated with particular haplogroups is more likely to be the dominant factor in survival or extinction.

German Dziebel said...

@ Tobus

I taught you a lesson on how to formulate the similarities between Amerindian and n/d population structure. You continue to revert to pseudoscientific objections. Ignored.

Linguists work with the time scales of human behavioral modernity and all known biological humans are behavioral humans as well so linguists work with the time scale of the human species. Your objection is again pseudoscientific and rooted in plain ignorance.

I don't have an arts major. I have two doctorates and that's the amount of science you obviously can't handle. Your creationist bias makes you believe that there's one and only one theoretical framework created by God. The data I referenced fully supports my inference that Amerindians are less ust ishim like than East Asians and more european like than other ust ishim like populations. Noise is what's happening in your ignorant head.

Ui wouldn't be pulled away from Amerindians if it's an East Asian west Eurasian hybrid as both of them are ultimately derived from a highly subdivided Amerindian population.

And I did present cases when heterozygosity increases with admixture. You haven't presented any evidence that the opposite is true. This is simply because you are a creationist blogger with no knowledge of population genetics. The relationship between admixture Fst and heterozygosity is very regular and supported by multiple studies.

Tobus said...

@Slumbery:
Thank you for the detailed answer. I am still not convinced.

And thank you for the rational discussion. One other aspect to consider is that the effect you are talking about is much stronger in the early stages of admixture. Consider the initial mating of 2 divergent individuals... at sites where one is AA and other BB then 100% of their offspring will be AB. But when these new AB individuals mate with each other, only 50% of their offspring (and future generations) will be AB at that site. Also, as time goes on either selection (which caused the BB in the first place) or drift will likely move this one way or the other so a few generations down the track we'd expect to see more AA or BB sites and less heterozygous AB, reducing the long term effect of these sites on heterozygosity.

That said, after considering your arguments I accept that during the admixture process when fully divergent individuals are still contributing there *could* be a spike in heterozygosity that *might* exceed both of the parents (if there are enough non-shared fixed sites), but once there are no "full blood" individuals left on either side, heterozygosity will quickly regress towards the mean of the contributing populations (in proportion to their contribution).

Are you sure that this is also true in the modern humans vs. Neanderthals or modern humans vs. Denisovans relation?

Yes, I'm sure. Neanderthals and Denisovans are highly inbred, and everyone knows that the offspring of an inbred and an outbred individual will have lower diversity than the offspring of two outbred individuals.

The reason for this is that we have far more alleles in common with Neanderthal/Denisovans ("N/D") than we have differences, and inbreeding will cause loss of heterozygosity all over the genome, not just the differences. Consider that at the point of Neanderthal/Sapiens divergence ~1 million years ago, our common homo ancestor had already accumulated ~3 million years worth of homo-specifc heterozygosity since diverging from chimps. Since 75% (3my/4my) of our heterozygous sites will be shared with N/D, we'd expect that ~75% of these heterozygous sites lost by N/D inbreeding will already be heterozygous in sapiens, and so reduce extant heterozygosity in an admixture event. With only 25% of sites having the potential to increase heterozygosity, and 75% lowering it, it makes sense that a N/D-Sapiens admixture would result in lower, not higher heterozygosity (from the sapiens point of view).

I'll add that this 75/25% guestimate isn't counting any earlier primate-common, or mammal-common, or vertebrate-common etc. heterozygous sites. Researchers have measured that only 0.12% of our total genome is different to that of Neanderthals, so using this figure we'd expect more like a 99.88% reduction and only a 0.12% increase.

Note that this would be different to admixture between two outbred populations where heterozygosity has been growing, not declining. In these cases the non-shared homozygous sites would be from de novo mutations within the 0.12% difference.0

Although it's not an ideal or conclusive test
case, I also note we don't see any obvious heterozygosity increase in UI or modern Eurasians from their Neanderthal admixture, nor any in Papuans who have close to 10% archaic admixture from 2 highly divergent sources.

Tobus said...

@Rokus:
Increased levels of heterozygosity can be measured within any mixed population, like the Caribbean, Mexico or Brazil.

The heterozygosity of these populations is only "increased" relative to the indigenous populations' heterozygosity - I think you'll find there is little or no increase relative to the colonising population.

Figure 2A of Verdu et. al. 2014 shows what I am talking about. Populations with recent and ongoing admixture like African-Americans, Mexicans and (native) North Americans show "increased" heterozygosity from their baseline but still at a level similar or lower than their higher contributor (although admittedly higher than I originally expected!). It also shows that populations with an admixed history like Middle Easterners and Cambodians have a heterozygosity within the range of their two contributing populations, the same thing that Laz. 2014 shows with the French being midway between their EEF and WHG ancestors. These data are consistent with the idea that resultant heterozygosity in admixed populations will tend towards the mean. While ongoing admixture might temporarily raise heterozygosity (due to there being complete sources of the divergent fixed alleles available which producing offspring who are heterozygous at 100% of these sites), this effect appears to disappear after the admixture process is complete (and only partial copies of the original fixed differences remain, so offspring are heterozygous at less than 50% of these sites).

Note that the African-American sample *does* score marginally higher than the highest African population (0.790 vs 0.786 - half a % increase) - the only example I've ever found, confirming that a slight increase during the admixture process is indeed possible. It's important to note that there's a significantly higher proportion of the high-heterozygosity population in the mix to start with in this case - so heterozygosity would start closer to the higher level and require a smaller number of divergent homozygous sites to go higher than if the low heterozygosity population was the primary component.

Bringing this back to the original point, if you're going to explain away UI's high heterozygosity by saying "archaic admixture" then you need to have an archaic population with heterozygosity at least as high as UI (if assuming recent and ongoing admixture), and something like twice as high if UI belongs to a discrete/stable population. That's assuming 50% or higher archaic admixture, at more realistic lower levels you would require proportionally higher heterozygosity in the archaic.


@German:
I intend to reply to your last post but it seems that we are cross-posting on this thread. I will wait for your response to my previous post (which is probably being published the same time as this one) and then respond to them both in one go. In the meantime you can read what I wrote to Slumbery and Rokus as it's all relevant to your position - and consider the examples where admixture *doesn't* increase heterozygosity that I gave from the Verdu and Laz papers.

Gregory76 said...

Terry says:

“Your own list shows three branches in SE Asia...The same holds for K2* which may or may not be a single haplogroup, perhaps one each in Sumatra and Sulawesi.”

I wasn't counting the ancestral haplogroup. But even if we do count it, it just breaks a tie.

“Turning to K2b. One branch, K2b1, is also specifically island SE Asia (and New Guinea).”

I already counted that. And don't forget that K2b2 (P) is not specifically SE Asian.

“And not to forget that K2a has a very strong presence in SE Asia although that is probably the result of later southward movement.”

You answer your own objection.

“But NOX could hardly be considered 'northern'.”

By “northern or western”, I mean Southwest Asian, Central Asian, North Eurasian (European or North Asian) or East Asian. Insofar as there is a finding of a given haplotype in any of these it fits comfortably with me theory that K2 originated in Southwest Asia and expanded to Central Asia, or originated in Central Asia, and in either case expanded from Central Asia to North Eurasia and East Asia, while only a finding of the haplotype in Southeast Asia, New Guinea, Melanesia or Australia would even count against my theory.

I earlier said:
"As for stragglers, it is very likely that large number of small bands of people became extinct while migrating, since it was easy for a small population to be wiped out by illness, war, natural disaster, etc."
Terry replied:
“True to some extent but in that case we should expect to find the odd pocket of surviving stragglers along the route. And a population migrating through a previously unoccupied region is unlikely to have remained 'a small population' for very long.”

I cannot easily accept either assumption, because death was so easy for early humans. So we should expect a high ratio of failures to successes regarding survival, and regarding increase in population.

Tobus said...

@German:
There are plenty of studies showing correlation between heterozygosty and admixture across species.

But none show the admired population having a heterozygosity higher than the that of parent population with the highest heterozygosity. See the Verdu paper.

Amerindians are too homozygous to allow for a mix between more heterozygous East Asians and West Eurasians.

You know very well that Amerindians spent some 10,000 years of bottleneck while crossing Beringia that lowered their heterozygosity. Don't pretend to be dumb, it suits you too well.

But it'd be good if you supported your speculations with an actual example of an admixed population with lower heterozygosity than both its parent populations.

Wordplay! "Not higher than both" does not mean "lower than both", it means in between them (ie "equal to or lower than one"). There are plenty of examples in the Verdu paper and the French from the Lazaridis paper.

I taught you a lesson on how to formulate the similarities between Amerindian and n/d population structure

By "formulate" you mean "make up". Your whole thesis was based on wordplay where "closest" was used to insinuate "close" when in fact N/D are the *furthest* population from Amerindians in term of heterozygosity. Now that we have an equally ancient sample that's far closer in both affinity and heterozygosity to Amerindians your spurious N/D association has been superseded.

Linguists work with the time scales of human behavioral modernity and all known biological humans are behavioral humans as well so linguists work with the time scale of the human species.

More word play! The oldest known language is about 5,000 years old. The oldest known genome is about 50,000 years old. No comparison.

The data I referenced fully supports my inference that Amerindians are less ust ishim like than East Asians

No, it fully supports the opposite - that none of these populations is more Ust-Ishim like than any of the others. Perhaps you should read Patterson again and refresh your memory on how the D-stat Z-scores work. If the data disagrees with your theoretical framework, change your theoretical framework, not the data.

Ui wouldn't be pulled away from Amerindians if it's an East Asian west Eurasian hybrid as both of them are ultimately derived from a highly subdivided Amerindian population.

I don't think you really understand how genetics works German. If a recently diverged population mixes with an already diverged population the result will be somewhere in between, not closer to the non-diverged descendants of the initial lineage. Draw a tree and compare the branch lengths.

...and BTW is this the same "highly divided Amerindians population" that is now "too homozygous to allow for a mix"?

And I did present cases when heterozygosity increases with admixture

Increases relative to the lower of the two contributors yes, but you've produced no examples yet where it's increased relative to the more heterozygous contributor.

You haven't presented any evidence that the opposite is true.

The very first time I raised this issue with you I presented the French as an example with heterozygosity midway between their EEF and WHG ancestors, and I repeated it earlier in this very thread. Do you have a memory/comprehension problem or do you just like lying when it suits your purposes?

Rokus said...

@TerryT,
Extremely unlikely to have sustained growth with an average of three surviving children per woman?? Actually, I was conservative on this. Take a look on the stats again, Niger has 7.6 births per woman on average, and the rest of Sub-Saharan Africa follows suit. Receiving all the help I don't think scarcity of natural resources will be a limiting factor soon, though maybe sustained worldwide economic crises will become one in Africa. New Zealand is a bad example, since growth in western countries fall well below the capacity of basic human resources, this has cultural reasons that are of no concern for prehistoric conditions.
However,for haplogroup drift and extinction we have to consider within-group growth differences. Examples abound! Within 13 generations my own surname grew from just one individual to 1,500, what equals 3.3 births per woman for each generation - and this only because births went down. By now there must be about 400 new generation kids around to continue procreation, what of course doesn't compare at all with Dutch national growth of "only" about 11 fold since the 16th century. Thus 'my' quite rare YDNA must have expanded at the cost of other lineages: drift and extinction within the borders of a single nation.
The current size of the Amish population could be explained by an average growth since 1737 (500) and the 19th century (3000) of 3.5 births per Amish woman. And still their natural resources are far from exhausted, since recently their birth rate increased to an average of over six births per woman. To have a picture of what this would look like: at this rate the Amish population would be twice the current American population in the second half of the 22nd century, and a staggering 60 billion one century later. Would they die of hunger? Maybe anywhere else but still not yet in the USA.
I agree such numbers are hard to imagine, though within populations such differential growth exists and will remain a feature in the future.
At least I hope you can see now that uneven growth doesn't respect majority haplogroups. Instead, one tiny haplogroup could pop up suddenly, and largely or completely replace all the others for any cultural reason.

Rokus said...

@Tobus,
I figure the accompanying text (Verdu et. al., 2014 ) is less ambiguous on the matter: "[...] heterozygosity levels among Pacific Northwest populations are substantially closer to those of Eurasian populations than to those of Central or South Americans. This result parallels the patterns observed in African Americans and Mexican Americans, two recently European-admixed populations in the Americas, who also showed inflated levels of genetic diversity compared to African and Native American source populations [...]. It is thus possible that admixture events following European contact might explain high genomic diversity in the Pacific Northwest populations in relation to Central and South Americans"

I am not sure you are aware that heterozygosity defines the pairwise variety of alleles inherited from each parent: heterozygosity does not occur in the gametes, and is fully the result of sexual reproduction. The actual number of heterogenous combinations should not be confused with the number of possible heterogenous recombinations, since heterozygosity is not individually acquired on site level. Heterozygosity is acquired by meiotic crossovers that involve large strings of DNA. Hence you have to think in centimorgans: "One centimorgan (cM) is equivalent to a 1% chance that a marker at one genetic locus will be separated from a marker at a second locus due to crossover per generation. In humans, 1 cM is equivalent to (on average) 1 million base pairs".
In other words, heterozygosity depends fully on the genetic differences between parents, not on their own individual heterozygosity on allel level. For the level of heterozygosity of offspring it is completely irrelevant how homozygous or heterozygous were the parents.

So UI's heterozygosity does *not* reflect any heterozygosity within prehistoric parent populations, that instead appear to have been predominantly homozygous compared with modern populations, and thus may be rather explained by admixture.

terryt said...

"it fits comfortably with me theory that K2 originated in Southwest Asia and expanded to Central Asia, or originated in Central Asia, and in either case expanded from Central Asia ... "

You seem to be searching for some single region from where all the haplogroup branches expanded, fully-formed, around the earth, like an ink blot through a blotter. That is what I call the 'Garden of Eden Syndrome. However it could probably be more accurately called the 'Noah's Ark Syndrome', diversity at origin.

Surely it is obvious that the reality was nowhere near that simple. To me it is far more likely that particular haplogroups have been able to expand fairly widely for one reason or another. Then founder effect, drift, bottlenecks, and even selection, have lowered haplogtype diversity in the various regions the original haplogroup has become isolated in. Some time later one or more of these descendant haplogroups has repeated the whole process. Always the edges get pruned unless the haplogroup enters some uninhabited region (and survives). We see credence for that view here:

http://eurogenes.blogspot.co.nz/2014/12/the-y-chromosome-tree-bursts-into-leaf.html

Even the most cursory glance at figure 3 from the original paper shows that the LT, NO, R split was long after the C, D. E, F split.

"I wasn't counting the ancestral haplogroup. But even if we do count it, it just breaks a tie".

Two problems here. Firstly the Sumatra/Sulawesi K2* is not an 'ancestral haplogroup'. It's just that its defining mutations are yet to be discovered. It is probably two separate haplogroups anyway. And secondly it does not 'just breaks a tie'. Even without it we have 2.5 SE Asian branches, 1 northern branch and 1 East Asian branch.

"You answer your own objection [K2a has a very strong presence in SE Asia although that is probably the result of later southward movement]".

Just one of NOX's surviving branches is found north of the modern Chinese border (as well as one extinct branch). K2a's southward movement was from within China, not from Central Asia.

"I cannot easily accept either assumption, because death was so easy for early humans".

Perhaps. We know of at least 3 lineages that have become extinct. There were obviously more though. The 3 can easily be seen as stragglers at the margin of a wider expansion rather than indicating source populations. They have been able to enter the steppe from either end of the Tibetan Plateau at favourable times, but being driven to extinction with the return of the cold. The oldest is Ust'-Ishim K2a at 45,000 years. Its looks to be an early entry onto the steppe from the east. Then extinction.

Kostenki C1b, at 37,000 years, belongs to one of the two C1 haplogroups that also became widespread. Surviving C1b is sporadically spread from northwest India to Timor. C1a is present at either end of northern Eurasia. C1c is confined to Australia. Kostenki C1b became extinct.

MA-1 at 25,000 years is R (I think R1). R1 is a good candidate for an origin in Central Asia, but it is the only one. Besides which K2b2 has become so widespread that it cannot provide any evidence for the origin of any other haplogroup.

terryt said...

A little clarification regarding Y-DNA K2:

"But even if we do count it, it just breaks a tie".

Far from being so. The 'one' northern haplogroup I claimed is actually made up of two half haplogroups, if you can imagine such a thing. K2a forms two branches: O, which can have originated no further north than northern China (certainly not Central Asia), and N, whose presence in the north is possibly relatively quite recent. That is 0.5 northern K2 haplogroups and 0.5 eastern so far.

K2b likewise forms two branches: K2b1 which can have originated nowhere but in southern Wallacea, and K2b2 which is northern but so widespread no obvious idea of its origin is possible. But I'll concede northern, even Central Asian. That makes 1 northern K2 haplogroup (0.5 plus 0.5), 0.5 eastern and 0.5 SE Asian.

K2c and K2d cannot be placed anywhere but in island SE Asia. That makes the ratio for K2 haplogrouops:

2.5 SE Asian (K2c, k2d and K2b1)

1 Central Eurasia (half a K2 haplogroup each in the form of N and K2b) and

0.5 East Asian (O).

That is nowhere near 'a tie'. And that is not even considering the widely separated K2* in both Sumatra and Sulawesi. Possibly making it 4.5 haplogroups for island SE Asia.

terryt said...

Apologies for a third comment here but I believe important points emerge from a study of the emerging phylogenies.

I said earlier, 'we should expect to find the odd pocket of surviving stragglers along the route'. To which Gregory76 said, 'I cannot easily accept either assumption, because death was so easy for early humans'. But the scenario I proposed is exactly what we see in the case of IJK. IJK splits off haplogroups all along the route from Near Africa, through South Asia, into East Asia, along the Southeast Asia peninsula and finally to Australia.

I agree K2b's deeper ancestry lies in southwest Asia. Several basal F-derived haplogroups have been identified there. For Example G looks to have originated somewhere near the Caucasus and certainly seems to be the oldest surviving haplogroup in those mountains:

http://dienekes.blogspot.co.nz/2014/11/paternal-lineages-and-languages-in.html

One branch of H (H2) is found primarily in SW Asia as is one branch of IJK (IJ). But that's it. Other F-derived haplogroups are South Asian (F1, F3, H1) or further east (H3, F2).

According to every phylogeny I've seen IJK splits into two branches: IJ and K-M9. K-M9 then splits in two: K1 (or LT) and K2. Neither L nor T are Central Asian. In fact they seem to be mainly South Asian. In any case the split between K1 and K2 probably did not take place in Central Asia. Apart from K2b2 all other K2 haplogroups are East or Southeast Asian, with K2b1a-P405 reaching New Guinea, and one haplogroup derived from that group (K2b1a-P60) even reaching Australia.

I agree haplogroups have probably become extinct but we have to rely on the evidence as we have it. It is no good making up ideas of extinction just to fit pre-existing beliefs.

terryt said...

"one tiny haplogroup could pop up suddenly, and largely or completely replace all the others for any cultural reason".

Generally technologies are shared regionally, or at least tribally, and so ultimately all haplogroups within a region share the expansion.

"Extremely unlikely to have sustained growth with an average of three surviving children per woman?? Actually, I was conservative on this. Take a look on the stats again, Niger has 7.6 births per woman on average, and the rest of Sub-Saharan Africa follows suit".

And all survive? You know as well as I do that the population growth in Africa is a recent phenomenon and cannot continue at its present rate. Long-term it cannot be sustained. You even admit:

"Receiving all the help I don't think scarcity of natural resources will be a limiting factor soon"

Yes. It is an artificial situation and scarcity of natural resources regularly leads to widespread famine in Africa.

" New Zealand is a bad example, since growth in western countries fall well below the capacity of basic human resources"

I was actually meaning the pre-European population. There is a huge amount of evidence that a collapse of resources led to huge population collapse after a few hundred years of population growth. That has presumably been the normality during human expansion.

"Within 13 generations my own surname grew from just one individual to 1,500, what equals 3.3 births per woman for each generation"

In a stable population such expansion would necessitate a considerable extinction of other surnames obviously. Human population is extremely unlikely to have been exponential since modern humans first appeared.

"of course doesn't compare at all with Dutch national growth of 'only' about 11 fold since the 16th century".

And you really believe that has been normal human population growth rate through all time?

"The current size of the Amish population could be explained by an average growth since 1737 (500) and the 19th century (3000) of 3.5 births per Amish woman. And still their natural resources are far from exhausted"

Yes. Because of vast improvements in agriculture. Such rapid improvement has only happened since plant and animal breeding techniques were improved over the last few hundred years. Before that famine and natural disasters regularly lowered population numbers.

"Would they die of hunger? Maybe anywhere else but still not yet in the USA".

Don't bet on it!

Tobus said...

@Rokus:
"heterozygosity levels among Pacific Northwest populations are substantially closer to those of Eurasian populations"

Exactly - it's "closer to", not "higher than", Eurasians. In all cases (bar the marginal increase of African-Americans noted above) the heterozygosity of admixed populations is somewhere in between the heterozygosity of the parent populations - it's not lower than the lowest parent nor is it higher than the highest parent. If you posit UI as an admixed population with one parent having extremely low heterozygosity (ie. Neanderthal/Denisovan level), then we'd have to expect the other parent to have heterozygosity at least as high as UI, if not significantly higher.

For the level of heterozygosity of offspring it is completely irrelevant how homozygous or heterozygous were the parents.

I think you are confusing yourself by mixing individual and population-wide scenarios. While an individual can be heterozygous at a site where both parents are homozygous for different alleles, these sites must *already be heterozygous* within the wider population for both alleles to be present in the two parents. Sites that aren't heterozygous in a population as a whole will *never* be heterozygous in the offspring of two individuals from that population ( de novo mutations excepted of course). So while it's possible to craft scenarios where an individual is heterozygous at different sites to their parents and claim there's no connection, the average heterozygosity of an individual's parents is actually *extremely* important in determining individual heterozygosity - it determines both the number of sites that *can* be heterozygous and the *probability* that each of these will be so in a single random sample.

So UI's heterozygosity does *not* reflect any heterozygosity within prehistoric parent populations

As I explained above this is simply not correct. UI's heterozygosity is a *direct result* of his prehistoric ancestors' heterozygosity. It can't really be any other way... think about it.

German Dziebel said...

@Tobus

"You know very well that Amerindians spent some 10,000 years of bottleneck while crossing Beringia that lowered their heterozygosity."

You can't use one interpretation (the Beringia crossing is just an idea, which is in need of a proof) as a proof of another interpretation. That's a typical creationist strategy: let's me proof my idea with a reference to a sacred text.

"But none show the admired population having a heterozygosity higher than the that of parent population with the highest heterozygosity."

UI is within the range of West Eurasian heterozygosity and WU is one of its parent populations. What's your problem?

"See the Verdu paper."

The Verdu paper in fact clearly shows spikes of heterozygosity in Old World-admixed populations of the New World compared to unadmixed New World populations.

"By "formulate" you mean "make up". Your whole thesis was based on wordplay where "closest" was used to insinuate "close" when in fact N/D are the *furthest* population from Amerindians in term of heterozygosity."

Like all creationists, you call logic "wordplay" of crafty scientists. Of course, Amerindians are modern humans. They are not D or N. But N/D are further from Africans than Amerindians are from N/D. This makes Amerindians closer to N/D than Africans are. Of course, a creationist abhors the thought of human descent from lower hominins, hence you're trying to sweep under the carpet the logical inference from the data, namely that some modern humans are closer to Eurasian hominins than others and that there's a cline linking all of modern humans into a single history of descent/speciation (hence the break between Amerindians and N/D) followed by migration of parts of the new species (hence the stronger similarity of Amerindians to the rest of humans than to N/D). Amerindians must have stayed in isolation from their D/N ancestors for much longer than the amount of time that has elapsed since humans reached Africa, hence the phased pattern of evolution from N/D to Amerindians.

"he oldest known language is about 5,000 years old. The oldest known genome is about 50,000 years old. No comparison."

No comparison is due to your ignorance. Linguistics is much older than genetics and it has generated pretty meaningful results so far. And since they are perfectly in line with population genetics, ancient DNA will only advance our knowledge of linguistic evolution.

"and BTW is this the same "highly divided Amerindians population" that is now "too homozygous to allow for a mix"?"

We're talking about two different kinds of measuring genetic diversity, stupid! A tutorial is available here: http://anthropogenesis.kinshipstudies.org/2014/12/the-best-kept-secret-in-populaton-genetics-or-truth-about-african-genetic-diversity/.

"The very first time I raised this issue with you I presented the French as an example with heterozygosity midway between their EEF and WHG ancestors, and I repeated it earlier in this very thread."

And who's midway in the case of UI? And French have ANE component as well, no?

Gregory76 said...

Terry said:
“You seem to be searching for some single region from where all the haplogroup branches expanded, fully-formed, around the earth, like an ink blot through a blotter.

No: I am quite comfortable with a theory which says that K2 did not give rise to K2c and K2d until I went to Southeast Asia/New Guinea, and though I think that K2a, K2b and Ust-Ishim originated somewhere in Southwest Asia, Central Asia or southern Siberia (or perhaps India), those, areas cover a large amount of territory, and added together covered even more. So I am not proposing that they all branched off in small area.

Terry continued:
“And secondly it does not 'just breaks a tie'. Even without it we have 2.5 SE Asian branches, 1 northern branch and 1 East Asian branch.”

No: That does not add up to 5 branches: rather, there are 2.5 branches that are cleary from SE Asia/New Guinea and 2.5 that are not (K2a, Ust-Ishim and K2b2):

K2c—Southeast Asia
K2d—Southeast Asia

Ust Ishim—North/Central Asia
K2a (NOX)—mostly northern, with its Southeast Asian branch arriving there from East Asia

K2b (MPS)—two sub-branches:
K2b1—Southeast Asia/New Guinea
K2b2 (P)—mostly northern or western distribution

So that is two branches in Southeast Asia, two mostly northern or western, and one with a sub-branch in each area. So we can no longer say that a clear majority of the branches of K2 originated in Southeast Asia.


Terry continued:
Just one of NOX's surviving branches is found north of the modern Chinese border (as well as one extinct branch). K2a's southward movement was from within China, not from Central Asia.

I never said K2a reached Southeast Asia directly from Central Asia: I assume it went through China first

I continued:
"I cannot easily accept either assumption, because death was so easy for early humans".

Terry replied:
“Perhaps. We know of at least 3 lineages that have become extinct.”

More than perhaps. Life was very hard for them and death was very easy.
People have accomplished some remarkable feats—but they are remarkable because they are so unusual.

Gregory76 said...

Terry said:
“That makes the ratio for K2 haplogrouops:

2.5 SE Asian (K2c, k2d and K2b1)

1 Central Eurasia (half a K2 haplogroup each in the form of N and K2b) and

0.5 East Asian (O).

That is nowhere near 'a tie'.”


Only haplogroups clearly confined or almost confined to SE Asia/New Guinea count against my hypothesis. So haplogroups in SW Asia, C Asia, N Asia, Europe and E Asia are alike in not counting against my hypothesis.
So the tally is still even:

Two branches of K2 and one of two sub-branches of K2b (MPS) in Southeast Asia/New Guinea:
K2c
K2d
K2b1

and two other branches of K2 and the other of the two sub-branches of K2b (MPS) not clearly confined or almost confined to Southeast Asia/New Guinea:

Ust Ishim—North/Central Asia
K2a (NOX)—mostly northern, with its Southeast Asian branch arriving there from East Asia
K2b2 (P)—mostly northern or western distribution

terryt said...

"Only haplogroups clearly confined or almost confined to SE Asia/New Guinea count against my hypothesis. So haplogroups in SW Asia, C Asia, N Asia, Europe and E Asia are alike in not counting against my hypothesis".

If you're going to rely on that approach you should not count any haplogroups not 'clearly confined or almost confined to' specific regions. Unless you're already committed to some theory.

"I am quite comfortable with a theory which says that K2 did not give rise to K2c and K2d until I went to Southeast Asia/New Guinea"

Neither K2c nor K2d have been found in New Guinea. They are confined to SE Asia. K2 of any sort did not reach New Guinea until K2b1 had formed. Two branches of K2b1 have been found there. K2b1b and K2b1c are both Wallacean. This makes it extremely likely that K2b itself must also have existed somewhere in or near SE Asia.

"I think that K2a, K2b and Ust-Ishim originated somewhere in Southwest Asia, Central Asia or southern Siber"

I think that is so unlikely as to be not worth considering. Ust'-Ishim is most closely connected to East Asian K2a and so is most likely to be an early northward expansion of that haplogroup.

" rather, there are 2.5 branches that are cleary from SE Asia/New Guinea and 2.5 that are not (K2a, Ust-Ishim and K2b2)"

Wrong. Ust'-Ishim is part of K2a and so we're left with 1.5 branches that are not SE Asian (K2a and K2b2).

"K2a (NOX)—mostly northern, with its Southeast Asian branch arriving there from East Asia"

On what grounds do you claim X arrived from East Asia rather than that K2 as a whole arrived in East Asia from South Asia? The latter makes more sense to me.

"we can no longer say that a clear majority of the branches of K2 originated in Southeast Asia".

That is so only if you fiddle the evidence.

"Ust Ishim—North/Central Asia
K2a (NOX)—mostly northern, with its Southeast Asian branch arriving there from East Asia"

Much as you wish it Ust'-Ishim is K2a, not a separate haplogroup. And to claim NOX is 'mostly northern' is to completely ignore the facts.

Tobus said...

@German:
You can't use one interpretation (the Beringia crossing is just an idea, which is in need of a proof) as a proof of another interpretation.

http://www.scientificamerican.com/article/first-americans-lived-on-bering-land-bridge-for-thousands-of-years/

My point isn't whether the mainstream scientific position is correct or false, it's that you *know* it's the mainstream scientific position but you ignore it and then use it's fabricated absence as evidence of your own position. This is dishonest. Relatively low Amerindian heterozygosity is entirely consistent with the mainstream view of human origins, please don't try to pretend otherwise.

UI is within the range of West Eurasian heterozygosity and WU is one of its parent populations.

We seem to agree - under a model where UI is the result of an archaic WEu/Amerindian admixture, the contributing archaic WEu population had a heterozygosity level similar to (or higher than) modern WEu populations... yes?

The Verdu paper in fact clearly shows spikes of heterozygosity in Old World-admixed populations of the New World compared to unadmixed New World populations.

Yes, and it clearly shows that the spikes in these admixed populations *don't* go higher than that of their Old World contributors. If UI is indeed an archaic New World/Old World admixture, this data strongly suggests that one of his parent populations had a heterozygosity level at least as high as him.

Like all creationists

I'm not a creationist German.

This makes Amerindians closer to N/D than Africans are

But it *doesn't* make Amerindians closer to N/D than to Africans. That's where your logical non sequitur is. You also have an underlying false assumption that heterozygosity between species is a definitive indicator of shared ancestry.

N/D and UI are similar in age, but UI is much closer in heterozygosity to Amerindians, and several orders of magnitude closer to Amerindians in terms of genetic affinity. The same data you use to suppose an N/D-like ancestor of Amerindians can be used to construct a *much* stronger case for a UI-like ancestor of Amerindians: "But UI is further from Africans than Amerindians are from UI. This makes Amerindians closer to UI than Africans are. Of course, a racist abhors the thought of human descent from black people hence you're trying to sweep under the carpet the logical inference from the data" etc. etc.

Linguistics is much older than genetics and it has generated pretty meaningful results so far.

Yes, but only back about 5,000 years - we need something than can tell us about things 100+kya in the past. Linguistics isn't going to do it - languages change too quickly and left no trace before writing.

A tutorial is available here: http://anthropogenesis.kinshipstudies.org/2014/12/the-best-kept-secret-in-populaton-genetics-or-truth-about-african-genetic-diversity/.

You can't use one interpretation (the divergence-then-admixture of North/South Amerindians is just an idea, which is in need of a proof) as a proof of another interpretation. That's a typical creationist strategy: let's me proof my idea with a reference to a sacred text.... and it's a sacred text you wrote yourself!!

And who's midway in the case of UI?

UI - you are still proposing UI as an admixed archaic WE/Amerindian population no?

And French have ANE component as well, no?

Which makes Stuttgart's higher heterozygosity even more contradictory to your "heterozygosity in admixed populations is always higher than in non-admixed ones" assumption.

terryt said...

Gregory76 will hopefully find Ebizur's comment at Dienekes post on Kostenki14 helpful:

" the Ust'-Ishim specimen does not belong to haplogroup X, but rather to a basal branch of haplogroup K2a(xNO'X). In other words, haplogroup NO-M214 and the Telugu haplogroup X are (slightly) more closely related to each other than either is related to the Ust'-Ishim specimen's Y-DNA. Likewise, haplogroup R1-M173 and haplogroup R2-M479 are more closely related to each other than either is related to the Mal'ta specimen's Y-DNA".

http://dienekes.blogspot.co.nz/2014/11/genome-of-kostenki-14-upper-paleolithic.html

In other words the Ust'-Ishm haplogroups you bang on about is an outlier to the more southern NOX, not a source haplogroups.

Rokus said...

Tobus, gradual differences in heterozygosity are just the penultimate stage of gene flow, before all marked differences disappear altogether. Hence you should not build conclusions on the admixture of adjacent populations since these were already admixing for centuries.

"Sites that aren't heterozygous in a population as a whole will *never* be heterozygous in the offspring of two individuals from that population"

Please don't change the subject, it's about admixture, that per definition involves various populations, not just one. Each of the parental populations may be fully homozygous for all sites and still they always produce heterozygous offspring.

Elsewhere I saw you comparing the levels of heterozygosity with a 'jar of marbles' of different colors. Well, I am not a teacher. I did my due explaining that heterozygosity defines the pairwise variety of alleles inherited from each parent; that heterozygosity is not individually acquired on site level but acquired by meiotic crossovers that involve large strings of DNA. Even mutation happens before of that. Marbles don't combine to heterozygous colours nor recombine, nor do jars contain subsets of marbles nor do they keep a record of their origin.

So far you just explained you don't understand heterozygosity at all and feel comfortable with it, but maybe you just don't want to understand Ust Ishim as a remarkably heterozygous and admixed individual. This insight should of course blow the mind of just anybody that ever chose to understand modern humans as a separate lineage.

terryt said...

@ Gregory76:

I presume you now accept there are four basal K2 haplogroups: K2a, K2b, K2c and K2d. You alos seem to accept both K2c and K2d are confined to SE Asia. That makes 50% of K2 haplogroups in SE Asia south of China and west of Wallace's Line even before we consider the other K2 haplogroups.

And that is ignoring the K2* listed as being present in the same region, possibly actually two separate haplogroups. But that would destroy your stance immediately and so I'l give you the benefit of any doubts you may have.

One branch of K2b is confined to SE Asia east and immediately round Wallace's Line. Therefore even without considering K2* we are above 50% SE Asian south of China and around Wallace's Line (62.5% in fact).

And let's not forget that K2a is especially common immediately to the north of these other K2(xK2b2) haplogroups, in China. Only at the northern and western geographic extremities of K2a's distribution does K2a move just outside China.

I hope that settles the matter. But I'm not that hopeful.

Tobus said...

@Rokus:
Hence you should not build conclusions on the admixture of adjacent populations since these were already admixing for centuries.

Firstly, do you have any real world evidence that admixture with the last century raises admixture to higher than both contributors? I note that three of the Verdu populations still contain "full-blood" members of each contributor, strongly suggesting that it does not. Secondly, it's extremely unlikely that that UI's population only experienced the supposed admixture within the preceding 100 years - if we use a very generous late OOA date of 50kya: - 100y / (50 - 40)ky = 1% probability that it happened in any particular 100y period... So no data to back up your initial premise and an extremely unlikely (but convenient!) circumstantial scenario.. instead of stretching reason to defend a preconceived assumption, I suggest it would be more scientific to let the data direct you to a new theory instead - however UI came about, at least one of his ancestors had high heterozygosity too.

Each of the parental populations may be fully homozygous for all sites and still they always produce heterozygous offspring.

Back to step one: this is true in theory but impossible in practice - no population can survive if it's fully homogenous, and no two populations can breed unless they share a significant amount of their heterozygosity. In any real world scenario, the bulk of heterozygosity difference is due to loss of common ancestral SNPs (ie bottleneck/inbreeding) and hence admixture will *lower* the heterozygosity of the more outbred population, not raise it. Think about it.

Elsewhere I saw you comparing the levels of heterozygosity with a 'jar of marbles' of different colors.

No, as I explained to Slumbery above this was an example of how averaging works, not of how heterozygosity works.

but maybe you just don't want to understand Ust Ishim as a remarkably heterozygous and admixed individual

Personally I see no reason to assume UI was admixed, but then again no reason to assume he wasn't either. My point is simply that *if* he is admixed, then the data would suggest that one of the contributing populations had a heterozygosity at least as high as him, if not higher. He didn't get his relatively high level of heterozygosity from admixture between two lowly heterozygous populations.

Gregory76 said...

I earlier said:
"Only haplogroups clearly confined or almost confined to SE Asia/New Guinea count against my hypothesis. So haplogroups in SW Asia, C Asia, N Asia, Europe and E Asia are alike in not counting against my hypothesis".

Terry replied:
If you're going to rely on that approach you should not count any haplogroups not 'clearly confined or almost confined to' specific regions. Unless you're already committed to some theory.

I reply:
Of course I am committed to a theory, and so are you. My point is that the evidence from the subgroups of K2 is not strong enough to decide the question in favor of your theory.

I said earlier:
"I think that K2a, K2b and Ust-Ishim originated somewhere in Southwest Asia, Central Asia or southern Siberia"

Terry replied:
I think that is so unlikely as to be not worth considering.

I reply:
That is because you are committed to your theory. But the strongest evidence for it is the distribution of the K2 subgroups, and that is not so strong as to make the merits of my theory not worth considering, even if Ust-Ishim is part of K2a

Terry asks:
On what grounds do you claim X arrived from East Asia

I reply:
I don't. According to Wikipedia, it is a very rare lineage from India. I assume it arrived there from SW Asia, while the NO branch headed north.

Terry said:
And to claim NOX is 'mostly northern' is to completely ignore the facts.

I reply:
N is heavily northern, and if O originated in East Asia then it, too, is one of the northern or western lineage. And X is not southeast Asian.

terryt said...

"if O originated in East Asia then it, too, is one of the northern or western lineage".

For how long has 'East Asia' been part of 'Northern Asia'? I think you need to look at a map. Perhaps you live in the USA as I have always found people there have little knowledge of geography outside their own region.

"My point is that the evidence from the subgroups of K2 is not strong enough to decide the question in favor of your theory".

It is sufficient to make your theory very untenable though, as in:

"the strongest evidence for it [Terry's theory] is the distribution of the K2 subgroups, and that is not so strong as to make the merits of my theory not worth considering, even if Ust-Ishim is part of K2a"

But it makes your theory extremely unlikely to be correct.

"And X is not southeast Asian".

By no stretch of the imagination is it 'Northern Asian'. In fact it is definitely 'South Asian'. Perhaps, once more, you are not aware that India and SW China share a border.

"I assume it arrived there from SW Asia, while the NO branch headed north".

How can X have 'arrived there from SW Asia' when it has no near relations at all in that region? Every one of its relations, except for K2b2, is centred on Southeast or East Asia. X's nearest SW Asian relation is IJ, some considerable genetic distance from it.

Rokus said...

@Tobus,
'In any real world scenario, the bulk of heterozygosity difference is due to loss of common ancestral SNPs (ie bottleneck/inbreeding) and hence admixture will *lower* the heterozygosity of the more outbred population, not raise it. Think about it.'
You have it all backwards. Bottlenecks and inbreeding decreases heterozygosity in a population, and increases homozygosity! Please read some basic textbooks on the subject before you continue this conversation.

'an example of how averaging works'
Not on heterozygosity! What you don't know is that gametes for sexual reproduction are haploid cells, where heterozygosity simply doesn't apply. Heterozygosity only applies to diploid cells, where allelles of both parents come together and thus pair up into either homozygous or heterozygous combinations.

Only in the special case of adjacent populations, that already share a fixed set of allelles, admixture would be equally likely to recombine another, smaller set of variable allelles either into homozygous or heterozygous basepairs and thus have some kind of 'averaging' effect. Actually, intensive gene flow caused most modern Eurasian populations to behave as adjacent populations. Still there are marked differences between (heterozygous) urban populations and rural populations. Also, it has been attested that low levels of African admixture tend to boost eg. Mexican heterozygosity.

'this is true in theory but impossible in practice'
All theory must be wrong. This is indeed the point of view of a creationist. Coincidence?

'My point is simply that *if* he is admixed, then the data would suggest that one of the contributing populations had a heterozygosity at least as high as him, if not higher.'
Hilarious! No need at all. However, I don't exclude the possibility that Ust Ishim was transitional between an earlier admixed population having an even higher heterozygosity, and modern humans of moderate heterozygosity. Bottom line is such highly heterozygous forebears of Ust Ishim where all plain hybrids, whose parent populations in turn had each of them reduced levels of heterozygosity instead.

Tobus said...

@Rokus:
You have it all backwards. Bottlenecks and inbreeding decreases heterozygosity in a population

You say "backwards" but then repeat my point exactly - bottlenecks and inbreeding reduce heterozygosity. What you don't seem to grasp is what this means in terms of admixture.

Consider that 1 million years ago there was *no* difference between AMH and Neanderthals, we were a single species and had the exact same level of heterozygosity - a level that had accumulated over some 3.5 billion years, with roughly 3 million years worth of it accumulating since our divergence from chimps alone. Only about 1 million years worth of our heterozygosity is uniquely our own, the rest (at least 75 and maybe even 99+%) is inherited from our common ancestor with Neanderthals.

As you state above in no uncertain terms, inbreeding lowers heterozygosity, and Neanderthals are *extremely* inbred with *extremely* low heterozygosity. It seems evident that the vast bulk of the difference in AMH/Neanderthal heterozygosity is due to a *loss of heterozygosity* in Neanderthals. Remembering that AMH and Neanderthals share well over 75% (like 99% even!) of their DNA, we'd expect 75% (99% even!) of this lost heterozygosity to be heterozygosity accumulated before the divergence from AMH. In terms of admixture, this means that virtually all of the homozygous sites in Neanderthals will either be ancestral, or will be already heterozygous in AMH.

Not on heterozygosity!

Yes on heterozygosity - for the most part. Sites that are already heterozygous in both populations will be averaged, as will sites that are already heterozygous in only one of the populations - run through a few examples on paper and you'll see what I'm saying. What happens to sites that are 50% in each population? What about 75% in one, 25% in the other? What about 50%/0%? Etc.

The only sites that don't get averaged are those that are homozygous in both populations, and then only those where each populations is fixed for a different alleles. These will be heterozygous in 50% of the resultant population (or whatever %age the admixture proportion is). In practice there are very few of these and they are far outweighed by the averaging effect in the rest of the genome. The fact that the resulting heterozygosity in roughly averaged (as evidenced by the Verdu paper) makes sense when you consider the vast bulk of our DNA is shared - even archaics like Neanderthals only differ to us in perhaps some 1% of their entire genome - and that the differences in heterozygosity between populations is primarily due to the *loss* of previously heterozygous sites due to bottlenecks etc. (otherwise, since we're all the same age with roughly similar demographics until the Neolithic, we'd all have pretty much the same heterozygosity level).

low levels of African admixture tend to boost eg. Mexican heterozygosity.

How many times do I need to say this - the "boost" you see is still lower than starting heterozygosity of the higher-heterozygous population. Perhaps you need look at it from the other way round - does mixing with Mexicans raise or lower the heterozygosity of Africans? Did Europeans gain or lose heterozygosity after mixing with North Americans?

creationist

No need to start insulting me Rokus. We're both intelligent men and I'm sure we can sort fact from fiction here using logic and reason - if you don't have a rational argument to make then it's probably because you are wrong.

Bottom line is such highly heterozygous forebears of Ust Ishim where all plain hybrids, whose parent populations in turn had each of them reduced levels of heterozygosity instead.

That's just passing the buck - the logic doesn't only apply to UI, but equally to all human populations. Do you have a real world example of what you are suggesting or do you not?


Rokus said...

'No need to start insulting me Rokus.'
Sure it is. I don't trust your rejection of theory and logic. You don't even care to sustain your rejection of the most obvious by citing scientific investigation. German already cited a useful link that proves you wrong, if you had only cared to read it:

Rudan et al. - Quantifying the increase in average human heterozygosity due to urbanisation (2008)
http://dx.doi.org/10.1038/ejhg.2008.48

'The human population is undergoing a major transition from a historical metapopulation structure of relatively isolated small communities to an outbred structure. This process is predicted to increase average individual genome-wide heterozygosity (h)'

'the most 'outbred' individual had an additional 21% of his/her genome heterozygous in comparison to the most ‘inbred’ individual.'

'We show that levels of h are increasing through the processes of urbanisation, isolate break-up and admixture.'

Highly heterozygous populations will always tend to loose heterozygosity over time, only having a dependency on size. Admixture between isolated populations will aways tend to increase heterozygosity. Both processes have been at work for tens of thousands of years. In the past populations were smaller, more isolated and more homozygous, and still much more different one from the other. Fluctuations of heterozygosity thus must have been much more dramatic in the past than nowadays. Ust Ishim is a valid testimony of these early times of Homo Sapiens as a new hybrid species.

German Dziebel said...

@Tobus

"We seem to agree - under a model where UI is the result of an archaic WEu/Amerindian admixture, the contributing archaic WEu population had a heterozygosity level similar to (or higher than) modern WEu populations... yes?"

East Asian-West Eurasian admixture, I wrote, not West Eurasian-Amerindian admixture. MA-1 is WE/Amerindian admixture. UI is East Asian/WE admixture. You still need to provide evidence whereby composite heterozygosity never exceeds the heterozygosity of the largest contributing population. It may or it may not depending on how divergent the populations are, how large, etc., how long the admixture has been in place, etc. (In Verdu Fig. 2, there's a "U.S. admixed" population that may exceed the heterozygosity of the most heterozygous African population in the sample.) The important aspects are that a) an admixed population has elevated heterozygosities compared to an admixed one, so Amerindians would be like UI if they were a mix of West Eurasians and East Asians; b) UI still has Eurasian level heterozygosity, which is not what's expected if it were a mix of Africans ("the first occupants of Eurasia" = Africans) and Neandertals. We would expect it to be higher than that of modern Africans.

"Yes, and it clearly shows that the spikes in these admixed populations *don't* go higher than that of their Old World contributors. If UI is indeed an archaic New World/Old World admixture, this data strongly suggests that one of his parent populations had a heterozygosity level at least as high as him."

You are just confirming what I proposed: UI is a mix of WEu (a population with a similar heterozygosity to UI) and East Asians. The contributing WEu population may have been pre-admixed in its own right.

"My point isn't whether the mainstream scientific position is correct or false, it's that you *know* it's the mainstream scientific position but you ignore it and then use it's fabricated absence as evidence of your own position. This is dishonest. Relatively low Amerindian heterozygosity is entirely consistent with the mainstream view of human origins, please don't try to pretend otherwise."

This is exactly how myths are told. Amerindians split from East Asians (heterozygosity went down), then absorbed West Eurasians (heterozygosity went up), then got locked in Beringia (heterozygosity dropped again, this time from a West Eurasian high to being lower than East Asians and as such the closest to Eurasian hominins). This is turtles all the way down, just to please the creationist who still works with a Bible-like chronology for the New World. This is not science but to a creationist like you it very well may appear to be such. Creationists call themselves scientists, too, you know.

"I'm not a creationist German."

You are like the black white supremacist from a Dave Chapelle's skit. Scientific on the outside, creationist on the inside.

German Dziebel said...

@Tobus (contd.)

""But it *doesn't* make Amerindians closer to N/D than to Africans. That's where your logical non sequitur is. You also have an underlying false assumption that heterozygosity between species is a definitive indicator of shared ancestry."

I don't have any underlying false assumptions. N/D shares ancestry with Amerindians - it's firmly established by other means than homozygosity studies. It's interpreted as admixture but the heterozygosity cline shows that it's more likely ancient Eurasian and modern Amerindian population structure (hence, common descent) that accounts for this shared ancestry.

"Yes, but only back about 5,000 years - we need something than can tell us about things 100+kya in the past. Linguistics isn't going to do it - languages change too quickly and left no trace before writing."

Outspoken ignorance is one of the defining feature of a creationist. 6,000+ languages of the world have been meticulously classified into genetic units such as stocks and families. The relative diversity of linguistic stocks and families around the world is well known and it's those patterns that exceed the 5,000-10,000 year mark and go all the way to the formation of the current linguistic diversity landscape (under the single language origin assumption). 2/3 of that diversity (excluding PNG) is in the Americas. Africa and Europe are the least diverse continents linguistically. This pattern matches very nicely with the genetic patterns of intergroup and intergroup diversities. And since language is a defining aspect of modern humans, the source of modern humans must be extra-African, with the excess of African intragroup diversity coming from admixture (with archaic hominins). It's all very simple! Try to comprehend it while paying your marbles game."

terryt said...

"You still need to provide evidence whereby composite heterozygosity never exceeds the heterozygosity of the largest contributing population".

In any mixing of two populations the resulting heterozygosity will be somewhere between the two contributing populations. The example Rokus gave of modern heterozygosity is the result of mixing of multiple populations. Such multiple mixing would have been fairly rare in the past, especially during the Paleolithic. This is the explanation for:

"there's a 'U.S. admixed' population that may exceed the heterozygosity of the most heterozygous African population in the sample.)"

Tobus said...

@Rokus:
I don't trust your rejection of theory and logic

I don't reject theory and logic, I use logic to reject theories that are contraindicated by the data. A theory that doesn't come from the data is little more than make believe.

You don't even care to sustain your rejection of the most obvious....

"most obvious"? When the only early AMH sample we have has a certain level of heterozygosity the "most obvious" position is that that's just the level of heterozygosity that early AMH had - Stuttgart had even higher heterozygosity. Explaining it away as archaic admixture between two pops with low heterozygosity seems very unobvious and arbitrary to me.

..by citing scientific investigation

I've already cited Lazaridis 2014 and Verdu 2014. You can also add Friedlaender 2008 to the list - it shows that Polynesians are a mix of indigenous Taiwanese and Melanesians and have a heterozygosity mid-way between them.

German already cited a useful link that proves you wrong

On the contrary it confirms what I'm saying - the "admixed" sample in Fig. 2 has heterozygosity midway between it's contributing "inbred" island populations and the "outbred" urban population.

'the most 'outbred' individual had an additional 21% of his/her genome heterozygous in comparison to the most ‘inbred’ individual.'

This just shows that inbreeding lowers heterozygosity, we all know that already... In UI's case what you seem to be claiming that admixture between two "inbred" groups will result in an "outbred"- like heterozygosity. This paper doesn't show that, it confirms that admixing with an "inbred" population will *lower* the heterozygosity of an "outbred" population.

Ust Ishim is a valid testimony of these early times of Homo Sapiens as a new hybrid species

I see no data to suggest this and have no a priori reason to assume it, but *if* it's true, the data would indicate that at least one of the contributing populations had a (natural) heterozygosity as high, if not higher, than UI has.

Tobus said...

@German:
East Asian-West Eurasian admixture, I wrote

I thought East Asians didn't exist until the last 10,000 years in your theory?

You still need to provide evidence whereby composite heterozygosity never exceeds the heterozygosity of the largest contributing population.

I've provided plenty of examples where it doesn't - whereas *you* are yet to provide an example where it does. Asking me to prove it "never" happens is like asking me to prove that pigs "never" fly - knowing that it's never been recorded is enough to make my point, it's up to you to prove the undocumented case if you're going to insist upon it.

In Verdu Fig. 2, there's a "U.S. admixed" population that may exceed the heterozygosity of the most heterozygous African population in the sample.

Yes, as I pointed out when I first provided it. The ASW sample in question is predominately African, is still in the admixture process and the heterozygosity is raised by less than 5%. If you are suggesting UI is a similar case then this suggests that UI is more archaic than sapiens, the admixture was recent and ongoing, and that the archaic parent had a heterozygosity around 90+% as high as UI's.

The important aspects are that a) an admixed population has elevated heterozygosities compared to an [un]admixed one

Wrong, every single discrete admixed population I've found data for has a heterozygosity *lower* than it's highest parent (Middle Easterners lower than Africans, French lower than Stuttgart, Polynesians slower than Melanesians etc.). It's only non-discrete "admixing" populations (ie where there are still "full blood" individuals) that seem to show a similar level to the more heterozygous population (eg. Mexicans and North Amerindians similar to Europeans), and in only one example have I seen a (very marginal) increase over the more heterozygous parent. Admixed populations typically show a reduction in heterozygosity from their more heterozygous parent population, and an increase only from the point of view of their less heterozygous one - a commonly overlooked fact.

so Amerindians would be like UI if they were a mix of West Eurasians and East Asians

No, if they didn't undergo a recent bottleneck they'd be somewhere in between.

UI still has Eurasian level heterozygosity

You mean Eurasians still have UI-level heterozygosity... time goes forward.

which is not what's expected if it were a mix of Africans.. and Neanderthals. We would expect it to be higher than that of modern Africans.

No again. OOA Bottleneck ignored, we'd expect admixture between an outbred population like Africans and an inbred one like Neanderthals to *lower* the African heterozygosity, not increase it.... or are you suggesting that inbreeding is good for diversity?

UI is a mix of WEu (a population with a similar heterozygosity to UI)

*could be*, not *is*.

The contributing WEu population may have been pre-admixed in its own right.

That's just passing the buck - one of the populations contributing to the "pre-admixture" would still have had to had high heterozygosity.

This is turtles all the way down

... and yet is so much more consistent with the data than your alternative proposal.

You are like the black white supremacist from a Dave Chapelle's skit

And you are like that little kid in primary school who can't bear to admit when he's wrong so he starts calling people names instead. Let's just stick to facts and logic no?

Tobus said...

@German (cont):
I don't have any underlying false assumptions.

Yes you do - that heterozygosity necessarily equates to genetic affinity. Where is the empirical evidence that the population with the closest heterozygosity to N/D must therefore be their closest descendant? It seems counter-intuitive since UI has the closest genetic affinity to Neanderthals and yet is the furthest Eurasian from them in terms of heterozygosity... and interesting that Papuans are *not* the closest to Denisovans in heterozygosity but they *are* the closest to them genetically.

N/D shares ancestry with Amerindians

.. and even more ancestry with East Asians and UI.

It's interpreted as admixture but the heterozygosity cline

The only heterozygosity cline is from Africans down to Amerindians, N/D are a clear outgroup to Sapiens, the distance form them to the closet Sapiens being larger than the entire Sapiens range.

*If* heterozygosity *were* a determinative indicator of ancestry you're missing the obvious anyway - UI is much closer to Amerindians than N/D is in terms of heterozygosity *AND* is much closer in terms of genetic affinity as well. If you're looking for Amerindians closest ancestor at 40kya, UI is your man, N/D aren't even close.

it's those patterns that exceed the 5,000-10,000 year mark

Yes, but we need 50-100kya if not earlier.

Africa and Europe are the least diverse continents linguistically

That depends on a narrow definition of "diversity" - African has far more languages than America, they are just more related to each other (so not split into as many "families"). The pattern of language diversity seems to imitate the pattern of genetic diversity somewhat - Africa has a lot of diversity that is spread widely, whereas America has less diversity but it's in distinct isolated pockets. In any case, no one's got a clue what languages were like in UI's time or before, so it's irrelevant to this discussion.

Gregory76 said...

I said earlier:
"if O originated in East Asia then it, too, is one of the northern or western lineage".

Terry replied:
For how long has 'East Asia' been part of 'Northern Asia'? I think you need to look at a map. Perhaps you live in the USA as I have always found people there have little knowledge of geography outside their own region.

I reply:
My knowledge of geography is just fine, thank you. You need to read what I say more carefully. As I said earlier:
"Only haplogroups clearly confined or almost confined to SE Asia/New Guinea count against my hypothesis. So haplogroups in SW Asia, C Asia, N Asia, Europe and E Asia are alike in not counting against my hypothesis".
So, as far as this debate is concerned, the regions in question fall into only two sets, those that count against my hypothesis (SE Asia and New Guinea) and those that do not count against it (SW Asia, C Asia, N Asia, Europe and E Asia). For short, we can call the latter set of regions “Northern and Western Regions” (East Asia is north of the Himalayas, as a map will show you).


I said earlier:
"My point is that the evidence from the subgroups of K2 is not strong enough to decide the question in favor of your theory".

Terry replied:
“It is sufficient to make your theory very untenable though, as in:

[quoting me]'the strongest evidence for it [Terry's theory] is the distribution of the K2 subgroups, and that is not so strong as to make the merits of my theory not worth considering, even if Ust-Ishim is part of K2a'

But it makes your theory extremely unlikely to be correct.”

I reply:
No it does not.
Even if all that mattered were the number of haplotypes, it would not give a big advantage to your theory, because, if Ust-Ishm is NOX, the tally is now:

Two branches of K2 and one of two sub-branches of K2b (MPS) in Southeast Asia/New Guinea:
K2c, K2d and K2b1

and one other branch of K2 and the other of the two sub-branches of K2b (MPS) not clearly confined or almost confined to Southeast Asia/New Guinea:
K2a (NOX)—mostly northern, with its Southeast Asian branch arriving there from East Asia
K2b2 (P)—mostly northern or western distribution

2 ½ to 1 ½ is not a big advantage.

And the number of haplotypes in a region is not all the matters, nor must it always be decisive. While you may rest your theory solely on haplotype counting, I do not do so with mine. I have other weighty regions, such as the difficulty of migration and the need for NO to have originated in the far north to develop the Mongoloid haplotypes.


I said earlier:
"And X is not southeast Asian".

Terry replied:
“By no stretch of the imagination is it 'Northern Asian'. “

I reply:
I never said it was. (As my next comment indicated, see below.) The point is that it is not from SE Asia, nor from New Guinea, and only those two regions count against my theory (see above).

I said earlier:
"I assume it arrived there from SW Asia, while the NO branch headed north".

Terry asks:
How can X have 'arrived there from SW Asia' when it has no near relations at all in that region? Every one of its relations, except for K2b2, is centred on Southeast or East Asia.

I reply:
X is a branch of NOX (K2a), the other branch being NO, consisting of N, which is mostly in North Asia, and O, which probably originated in East Asia. With X being in South Asia, the ancestor of X and NO probably originated somewhere in between, such as Central Asia, Southwest Asia or Southeast Asia, and then divided into a northern branch (NO) and a southern branch (X). Those facts are compatible with either of our theories. And as I have said, there are strong reasons for supposing it originated in Southwest Asia or Central Asia rather than Southeast Asia.

Rokus said...

@Tobus,
'the "most obvious" position is that that's just the level of heterozygosity that early AMH had - Stuttgart had even higher heterozygosity.'
UI nor Stuttgart heterozygosity imply African genetic affinity, so non-African heterozygosity must have a different origin. Hence, African heterozygosity can't be equalled to early AMH. Moreover, Stuttgart didn't reach the UI level of heterozygosity, that was higher. Lazaridis rejected Stuttgart's high GATK value and conformed to the corrected mlRho value.

Fu et al. (2014), using the GATK method:
'The results (Table S12.1 and Figure S12.1) indicate that Ust'-Ishim has on average 8.17 heterozygous sites every 10,000 basepairs, a value that is ~4-fold higher than the heterozygosity in the archaic genomes of the Altai Neandertal and Denisovan individuals.'
'[...] Ust'-Ishim heterozygosity is slightly higher than that of other non-Africans (range 5.48-7.77)'

Lazaridis et al. (2013), using the software package mlRho (since here 'GATK estimates do not correct for the genotyping error that occurs in the context of low coverage, and hence may produce artifactual overestimates of heterozygosity'):
'Stuttgart has 7.42 heterozygous sites per 10,000 base pairs.'
'[...] only slightly higher than the most diverse present-day non-African in the panel'

Also UI's mlRho results are higher than Stuttgart: 7.67.

It's not surprising that Stuttgart measured high heterozygosity, since this specimen was admixed. Lazaridis: 'there is direct evidence that Stuttgart has European hunter-gatherer ancestry (Fig. 1B and Table S13.1).'
Actually, 'Stuttgart, fit as admixed with 47-79% Loschbour-related ancestry and 21-53% “basal Eurasian” ancestry'

I will ignore your further comments on averaging heterozygosity. I will not play your ridiculous game, you just don't know what you're talking about.

terryt said...

"Yes you do - that heterozygosity necessarily equates to genetic affinity".

I can't understand at all why German thinks that.

"there are strong reasons for supposing [Y-DNA X] originated in Southwest Asia or Central Asia rather than Southeast Asia".

Example of a 'strong reason'?

"Only haplogroups clearly confined or almost confined to SE Asia/New Guinea count against my hypothesis. So haplogroups in SW Asia, C Asia, N Asia, Europe and E Asia are alike in not counting against my hypothesis".

That's shifting geography to make it fit your theory. Why insist SE Asia is a separate region from East Asia when today the genetics of the two regions are very closely related? The two regions form a distinct group, separate from all the others you mention.

"those that do not count against it (SW Asia, C Asia, N Asia, Europe and E Asia)".

A lumping together of distinct and relatively isolated regions through which just two haplogroup branches are spread. Why is East Asia included anyway?

"Two branches of K2 and one of two sub-branches of K2b (MPS) in Southeast Asia/New Guinea:
K2c, K2d and K2b1"

Yes. Of course you're deliberately ignoring the obviously separate K* of Sumatra and Sulawesi, but I'll let that pass.

"K2a (NOX)—mostly northern, with its Southeast Asian branch arriving there from East Asia"

Only if you ignore the obviously ancient and close geographic connection between East and Southeast Asia. Far more connected than is East and North Asia. And you're ignoring South Asian X. K2a cannot realistically be called 'mostly northern'.

"2 ½ to 1 ½ is not a big advantage".

Including K* and 1/2 of K2a gives us three and a half Southeast and East Asia along with two half haplogroups for everywhere else. (half of K2b and half of K2a). That is quite a large advantage.

"such as the difficulty of migration and the need for NO to have originated in the far north to develop the Mongoloid haplotypes".

The Mongoloid phenotype probably developed as much through altitude as through latitude. The EDAR370a Mutation was suggested as originating somewhere near the Chin Ling Mountains of Northern China. I would guess NO accrued the phenotype somewhere nearby.

" The point is that it is not from SE Asia, nor from New Guinea, and only those two regions count against my theory (see above)".

Again you are ignoring contiguous geographic regions that we know have been genetically connected for thousands of years. South Asia shares a modern border with Southeast Asia but you seem obsessed with the idea that Paleolithic humans were confined within modern political borders. I can't understand how you can justify that concept.

"X is a branch of NOX (K2a), the other branch being NO, consisting of N, which is mostly in North Asia, and O, which probably originated in East Asia".

It is a really odds on bet that N developed somewhere near where O developed. And that is probably in East Asia, not in North Asia. Don't forget that N's expansion is comparatively recent although its separation from O is more ancient.

"the ancestor of X and NO probably originated somewhere in between"

Yes. East Asia is exactly between South Asia and North Asia.

German Dziebel said...

@Tobus

"I thought East Asians didn't exist until the last 10,000 years in your theory?"

My theory is rooted in a necessary inference from the fact that a 24,000 year old sample from (north)east Asia has no affinity to modern East Asians, while it has affinity to Amerindians. And Amerindians have affinity to East Asians that MA-1 does not have. So East Asians must have repopulated (north)east Asia in post-MA-1 times and this must have been from America. 12,000 years ago the passes got open, so this is when it must have happened. This interpretation fits all the other data.

But MA-1 shows that of course there were populations in East Asia prior to the second "Mongoloid" migration out of the Americas. MA-1 IS an East Asian population geographically and there were surely others as well (maybe "Papuan"/"Australoid" in a broad sense) that the subsequent back migrants from the Americas admixed with to create modern East Asians. So those ancient East Asians had admixed with West Eurasians to generate UI.

"I've provided plenty of examples where it doesn't - whereas *you* are yet to provide an example where it does."

I did. You even commented on it in your next paragraph. Are you senile?

"If you are suggesting UI is a similar case then this suggests that UI is more archaic than sapiens, the admixture was recent and ongoing, and that the archaic parent had a heterozygosity around 90+% as high as UI's."

Why archaic? UI heterozygosity falls within the modern West Eurasian range.

"No, if they didn't undergo a recent bottleneck they'd be somewhere in between."

Your bottleneck is just a feeble attempt to brush off the stubborn lack of agreement between facts and your theory. Out-of-Africa uses it in a similar context: there are no older African lineages outside of Africa (lack of facts to support the out-of-Africa theory) because they were destroyed by a bottleneck.

German Dziebel said...

@Tobus (contd.)

"The important aspects ...

Wrong>.."

How is it wrong if you yourself admit that it's right as in "an increase only from the point of view of their less heterozygous one". A dramatic increase, I must add. In any case, it's irrelevant since UI is within the bounds of West Eurasian heterozygosity consistent with your expectation of heterozygosity in admixed populations.

"t's only non-discrete "admixing" populations (ie where there are still "full blood" individuals) that seem to show a similar level to the more heterozygous population (eg. Mexicans and North Amerindians similar to Europeans), "

Again, you found the data to disprove yourself. Good, thanks.

"Neanderthals to *lower* the African heterozygosity, not increase it.... or are you suggesting that inbreeding is good for diversity?"

Any inbred population still carries some diversity unmatched by a population that had diverged from it long time ago. So they should add up. The lift from the outbred population's point of view would be minimal, though.

"heterozygosity necessarily equates to genetic affinity."

I specifically said multiple times that this is not the case. Affinity between N/D and Amerindians is proven by other means. The heterozygosity cline shows that this affinity is not due to admixture but to common descent from a highly substructured population. This is what people who debated admixture vs. substructure in Africa as models to explain "Neandertal affinities" of non-African genomes don't understand. In reality, it's substructure but not in Africa and it's common descent but again not in Africa.

"It seems counter-intuitive since UI has the closest genetic affinity to Neanderthals and yet is the furthest Eurasian from them in terms of heterozygosity"

That's precisely why UI must be admixed between 2 modern human, non-Amerindian populations. It's Neandertal affinity may not be due to admixture but to common descent that naturally both contributing post-New World populations had.

".. and even more ancestry with East Asians and UI."

No, less ancestry with East Asians than with Amerindians. yes, more ancestry with UI than with Amerindians but there's 45,000 years between Amerindians and UI so it's to be expected. Again, very importantly, Amerindians show the greatest proximity to MA-1 and to Neandertals than any other modern human population.

"The only heterozygosity cline is from Africans down to Amerindians, N/D are a clear outgroup to Sapiens, the distance form them to the closet Sapiens being larger than the entire Sapiens range."

The closest Sapiens to N/D, the outgroup to all modern humans, are Amerindians. So you again disproved yourself in one single sentence. African heterozygosity is a clear sign of population growth required to colonize Eurasia and Africa and admixture.

German Dziebel said...

@Tobus (contd.)

"UI is much closer to Amerindians than N/D is in terms of heterozygosity *AND* is much closer in terms of genetic affinity as well."

You are stating the obvious. Of course this is the case considering that Amerindians and UI are same species, while N/D are sister clades to all of modern humans. As a creationist, you are of course averse to the possibility of a hominin origin for man, hence you insist that humans originated from other humans.

"Yes, but we need 50-100kya if not earlier."

Any time horizon that's required to generate the extant diversity of the 6000 human languages. By definition. Diachronic patterns observed in a complete sample of necessity account for the origin of that sample, which is the origin of human language per se. Under an assumption of single origin, of course.

"The pattern of language diversity seems to imitate the pattern of genetic diversity somewhat"

Yes, this is what you learned from me. Good. Making progress.

"Africa has a lot of diversity that is spread widely, whereas America has less diversity but it's in distinct isolated pockets."

No, Africa has very little diversity that's spread across a much larger and denser population. America has a lot of diversity spread across a very small, sparsely distributed population.

Rokus said...

@TerryT
'In any mixing of two populations the resulting heterozygosity will be somewhere between the two contributing populations. The example Rokus gave of modern heterozygosity is the result of mixing of multiple populations.'
You contradict yourself. If such a genetic law would exist (and it doesn't!) that the resulting heterozygosity of two populations should be an avarage value (as if heterozygous basepairs are colored marles in a jar, they aren't!), then how on earth higher heterozygosity could ever occur with multiple populations? The mathematics for averaging should also apply for higher series, don't they? Likewise, the urban maxima of heterozygosity observed could never be the result of averaging the heterozygosity of adjacent populations! What's the average of 4 an 6? 5! What is the average of 1,2,3,4,5,6,7,8,9? Also 5! So where an urban heterozygosity of say 10 would come from? Other principles are involved. Please don't make things up, I thought you smarter than this.

Dienekes said...

In all cases (bar the marginal increase of African-Americans noted above) the heterozygosity of admixed populations is somewhere in between the heterozygosity of the parent populations - it's not lower than the lowest parent nor is it higher than the highest parent.

This is not always so. For example, if two populations have fixed differences at a locus, e.g., one is always A and the other is always C, then their heterozygosity at that locus is 0, but a mixed population will have a heterozygosity >0.


Tobus said...

@Dienekes:
For example, if two populations have fixed differences at a locus, e.g., one is always A and the other is always C, then their heterozygosity at that locus is 0, but a mixed population will have a heterozygosity >0.

You are correct, these sites will be 100% heterozygous in the 1st generation, 50% in the 2nd and then vary down from that due to drift, or more likely since one or both alleles were under extreme selection in the first place, selection. We'd expect sites like this to have a higher effect during the admixture process where fully divergent (ie "full blood") individuals of both parent populations exist, and a lesser effect as time goes on and the admixed population becomes discreet.

It's also important to consider that this is largely only talking about novel mutations *after* divergence which then became fixed - sites that were already heterozygous and became homozygous due to inbreeding/bottlenecks in one population (ie the vast majority Neanderthal/Denisovan sites) will already be heterozygous in the other contributor. This explains why mixing with an inbred will make your children half-inbred, not super-outbred.

The real-world question is how many of these novel fixed derived sites are there and do they outweigh the averaging effect of sites that are already heterozygous in either or both parent. From the data it would appear that as a rule, no, there aren't enough. ASW is the only exception I've found and the heterozygosity is only ~5% higher than the highest SSA. In this case the samples show a higher proportion of the African component and the admixture is still ongoing. It's also worth pointing out that Africans, as the deepest homo lineage, would be expected to have the highest number of private fixed alleles, making them most likely population we'd see this effect with (and AFAICT, they are the only one on record).

The rest of the examples, and there are lots of them, all show a resultant heterozygosity at or below the level of the highest contributor, with the "at" only applying to cases where the admixture is still ongoing.

Tobus said...

@Rokus:
UI nor Stuttgart heterozygosity imply African genetic affinity, so non-African heterozygosity must have a different origin. Hence, African heterozygosity can't be equalled to early AMH.

That's really bad logic Rokus, the conclusion simply doesn't follow from the premise - UI/Stuttgart heterozygosity doesn't *unimply* African affinity either.

Stuttgart didn't reach the UI level of heterozygosity, that was higher.

Confirmed - thank you for the correction! :)

I will ignore your further comments on averaging heterozygosity

Suit yourself.

then how on earth higher heterozygosity could ever occur with multiple populations?

The same way all genetic differences originate - de novo mutations.

Tobus said...

@German:
So those ancient East Asians had admixed with West Eurasians to generate UI.

In which case we'd expect UI to have greater affinity to the ancestors of ancient East Asians and West Eurasians than to the ancestors of your postulated ancient Amerindians, no?

Out-of-Africa uses it in a similar context: there are no older African lineages outside of Africa (lack of facts to support the out-of-Africa theory) because they were destroyed by a bottleneck.

The bottlenecks postulated in Out Of Africa are all correlated with significant drops in heterozygosity. They are bottlenecks suggested *by* the data, not created out of thin air to cover holes in it. That's the important difference between your theory and OOA, and why your theory is fringe and crackpot while OOA is universally accepted.

it's irrelevant since UI is within the bounds of West Eurasian heterozygosity consistent with your expectation of heterozygosity in admixed populations.

For the second time you agree that if UI is admixed then at least one of the contributing parents had heterozygosity higher than him - why are we still discussing this?

Any inbred population still carries some diversity unmatched by a population that had diverged from it long time ago

That "long time ago" is only a fraction of the period of time in which they accumulated heterozygosity as a single species - it's at most 25% of the time since homo divergence from chimps (using very generous 4mil chimp/homo and 1mil neand/sap dates, more like 6% if we use the 8mil/0.5mil extremes). The "add up" will be heavily skewed towards lost heterozygosity and the "minimal" increase would in fact be negative - a net *decrease* in heterozygosity.

Affinity between N/D and Amerindians is proven by other means

... and those means show that both both UI and Han are closer to Neanderthals than Amerindians are, and that Papuans/Melanesians are closer to Denisovans than Amerindians are.

That's precisely why UI must be admixed between 2 modern human, non-Amerindian populations.

Unless of course ancestral homo heterozygosity was similar to UI's level and N/D's low level is due to inbreeding... do you remember if the data says that N/D are inbred or not?

Amerindians show the greatest proximity to MA-1 and to Neandertals than any other modern human population.

Han Chinese show the greatest proximity to Neanderthals.

The closest Sapiens to N/D, the outgroup to all modern humans, are Amerindians

The closest Sapiens to N/D is Loschbour, and Amerindians are closer to every other Sapiens than they are to N/D, so there's no "cline", just a big gap for you to fill with your imagination.

Any time horizon that's required to generate the extant diversity of the 6000 human languages

So what language did UI speak? How was it related to the language MA-1 spoke? Or K14? Were any of these connected to Anzick's language? What about LB's, Loschbour's and Stuttgart's languages, was there any connection between them? How did any/all of these relate to the Neanderthal and Denisovan languages??

Use linguistics to suggest genetic relationships for times where we have no evidence of any language being spoken is futile. It's actually much more feasible to do it the other way round - if two ancient pops from the same time share genetics, they probably shared some language/culture too. There are many people trying to determine the source of Indo-European using genetics, since linguistics has already gone back as far as it can in that regard.

No, Africa has very little diversity

http://www.nationsonline.org/oneworld/languages.htm

terryt said...

Gregory. One more thing:

"East Asia is north of the Himalayas, as a map will show you"

The Himalayas are not the defining geographic boundary here. It is the whole Tibetan Plateau. Humans would have moved around the plateau, not through it. We therefore have four separate geographic regions marked by the plateau: East of it, North of it, west of it and south of it. Each of the four regions is reasonably genetically distinct with some overlap around the edges. East Asia is west of the plateau, not north of it.

Gregory76 said...

I said earlier:
"there are strong reasons for supposing [Y-DNA X] originated in Southwest Asia or Central Asia rather than Southeast Asia".

Terry asked:
Example of a 'strong reason'?

I reply:
First, as we both agree, the ancestors of these groups came from Africa, and the came from there by way of Southwest Asia. It is simpler to suppose that X arrived in South Asia by a direct route there from Southwest Asia, than to suppose that X migrated from SW Asia to South Asia and then migrated to SE Asia, and then migrated in the reverse direction back to South Asia.
Second, all of the reasons that favor a spreading of K2 descendants from SW Asia or C Asia rather than SE Asia,


Terry asked:
Why insist SE Asia is a separate region from East Asia when today the genetics of the two regions are very closely related? The two regions form a distinct group, separate from all the others you mention.

I reply:
A haplotype existing in SE Asia counts for your theory and against mine, but the haplotype existing in East Asia does not, since East Asia can just as easily be reached from Central Asia, as my theory as says, as from SE Asia, as your theory says. And E Asia is more like C Asia than SE Asia with regard to average temperatures, and both E Asia and C Asia are part of the Palearctic zoogeographic region,
whereas SE Asia is more like SC Asia than E Asia with regard to temperatures, and both are part of the Oriental zoogeographical region,



Terry said:
...you ignore the obviously ancient and close geographic connection between East and Southeast Asia. Far more connected than is East and North Asia.

I reply:
I addition to what I said above about E and SE Asia, I will add that the average temperature of E Asia is about as close to that of N Asia as it is to that of SE Asia, and that N Asia and E Asia are both part of the Palearctic zoogeographic region.



Terry said:
And you're ignoring South Asian X. K2a cannot realistically be called 'mostly northern'.

I reply:
The point is that X is not SE Asian, and, as I said above, its arrival in S Asia can be explained by a direct route from SW Asia, rather than going from SW Asia through S Asia to SE Asia and then back to S Asia.
Also, if Ust-Ishim is K2a, and is one two main branches of K2a, then that is all the more reason for thinking that K2a is northern.



Terry said:
Including K* and 1/2 of K2a gives us three and a half Southeast and East Asia along with two half haplogroups for everywhere else. (half of K2b and half of K2a). That is quite a large advantage.

I reply:
Again you are lumping E Asia and SE Asia together,



Terry said:
The Mongoloid phenotype probably developed as much through altitude as through latitude. The EDAR370a Mutation was suggested as originating somewhere near the Chin Ling Mountains of Northern China. I would guess NO accrued the phenotype somewhere nearby.

I reply:
I don't think North China is cold enough, being similar to NE Europe, and I'm not sure Tibet is cold enough: one climate maps seems to show that its temperature is that of the southern taiga and the Alps, while another shows it not even as cold as Mongolia.
As to the EDAR370a gene, from what I have been able to find out about it, is that is assocated with thicker hair and denser sweat glands, and the latter seems to be more useful in hot climates than cold ones.


Terry said:
“South Asia shares a modern border with Southeast Asia..”

I reply:
And South Asia shares a modern border with Southwest Asia.
And on your theory you must say that the ancestors of X came from SW Asia to S Asia, then went to SE Asia, and then reversed direction and came back to SC Asia, whereas I can say that the simply came from SW Asia to S Asia and stayed.


I said earlier:
"the ancestor of X and NO probably originated somewhere in between"

Terry said:
Yes. East Asia is exactly between South Asia and North Asia.

I reply:
And so is Central Asia and SW Asia.

terryt said...

"your theory you must say that the ancestors of X came from SW Asia to S Asia, then went to SE Asia, and then reversed direction and came back to SC Asia"

Why on earth is that impossible? Especially if I remind you that X has been discovered only in the northeast of India, not right across it.

"It is simpler to suppose that X arrived in South Asia by a direct route there from Southwest Asia, than to suppose that X migrated from SW Asia to South Asia and then migrated to SE Asia, and then migrated in the reverse direction back to South Asia".

I disagree. X lies within the K2 clade which, in spite of what you claim, is primarily a Southeast/East Asian bunch of haplogroups. Its route to SE/S Asia is debatable. I agree F as a whole must first have formed in SW Asia as several basal branches appear to be very much centred on the region. However I see no reason at all why all branches within F must have originated in in exactly the same place. Sounds a bit 'Garden of Eden' to me. That idea assumes humans were completely unable to move from SW Asia for some tens of thousands of years. A position I find very unlikely. I agree that neither X nor K2a need ever have actually been in SE Asia proper but must have originated somewhere very nearby. Either in northeast India or southwest China.

"the haplotype existing in East Asia does not, since East Asia can just as easily be reached from Central Asia, as my theory as says, as from SE Asia, as your theory says".

But when all of a particular haplogroup's related branches are present in SE Asia and none in Central Asia it makes SE Asia a more likely source.

"And E Asia is more like C Asia than SE Asia with regard to average temperatures"

Not true. Much of East Asia south of the Tsin Ling Mountains is tropical or subtropical. Central Asia also lies at a much higher altitude than does East Asia. Tibet is higher still. I can't imagine where you get the idea that China is part of the 'Palearctic zoogeographic region'. Reference?

"if Ust-Ishim is K2a, and is one two main branches of K2a, then that is all the more reason for thinking that K2a is northern".

Ust'-Ishim is one of three main branches of K2a. Don't forget that X is also part of that haplogroup. As a result K2a is spread from South Asia, through East Asia and out into Central Asia. Too widespread to be able to form any realistic idea of its region of origin through confining your perspective to just that haplogroup.

"Again you are lumping E Asia and SE Asia together"

With far more justification than your lumping of East and Central Asia.

"I don't think North China is cold enough, being similar to NE Europe"

So one minute China as a whole is part of the 'Palearctic zoogeographic region' and next minute even North China is not cold enough to explain the Mongoloid phenotype development. Tibet, with its extreme altitude, is probably far colder than most regions much further north.

"And so is Central Asia and SW Asia".

I don't think so [exactly between South Asia and North Asia]. You are ignoring the apparent fact that Tibet was settled relatively recently, long after most haplogroup branches had long been formed.

Gregory76 said...

I said earlier:
"It is simpler to suppose that X arrived in South Asia by a direct route there from Southwest Asia, than to suppose that X migrated from SW Asia to South Asia and then migrated to SE Asia, and then migrated in the reverse direction back to South Asia".
Terry said:
I disagree.
I answer:
You can disagree about much, but you cannot deny that the movement from SW Asia to S Asia is simpler than movement from SW Asia through S Asia to SE Asia and then back again to SC Asia.
Terry:
Why on earth is that impossible?
I reply:
Not impossible, just unlikely.

Terry continues:
X lies within the K2 clade which, in spite of what you claim, is primarily a Southeast/East Asian bunch of haplogroups.
I reply:
That K2 is primarily is a SE/E Asian bunch of haplogroups is what you must prove and to try to use it as part of the proof is to commit the Fallacy of Circular Reasoning.

Terry continues:
That idea assumes humans were completely unable to move from SW Asia for some tens of thousands of years.
I reply:
I don't assume that either. I assume that they moved from there to C Asia and S Asia, and from the former to N Asia and E Asia.

Terry said:
I can't imagine where you get the idea that China is part of the 'Palearctic zoogeographic region'. Reference?
I reply:
In addition to various textbooks I looked at in earlier years, there is the Wikipedia article on zoogeographic regions. They show Wallace's map, and a more recent map from the WWF, which moves the southern boundary of the Palearctic in China somewhat to the north, but still leaves most of China within the Palearctic: all of north China and some of south China.
Regarding the part of the comment that says: “Central Asia also lies at a much higher altitude than does East Asia”, I will point that it that is all the more reason for thinking that people moved from Central Asia to East Asia rather than the reverse, since they would be moving downhill rather than uphill!

Terry says:
Ust'-Ishim is one of three main branches of K2a.
I reply:
According to the comment of Ebizur (on Kostenki14), Ust-Ishim belongs to a branch of K2a that is outside NOX, making NO and X more closely related to each other than to Ust-Ishim. So we now have the branches of K2a as follows:

Ust-Ishim—found near the border of C Asia and N Asian
NOX
X—found only in Indi
NO
N—found mainly in N Asia, and not found south of E Asian
O—the only one found in SE Asia

And again, lest you claim that my grouping SW, C, N and E Asia together against SE Asia is arbitrary, the first four share the property of fitting comfortably with my theory, whereas only SE Asia does not.


I said earlier:
"I don't think North China is cold enough, "
Terry replies:
So one minute China as a whole is part of the 'Palearctic zoogeographic region' and next minute even North China is not cold enough to explain the Mongoloid phenotype development.
I reply:
Yes: East Asia is a mid-latitutude region. It follows from that both (1) that it is not cold enough for the Mongoloid phenotype to develop, and (2) that it is not tropical as SE Asia is.

Terry said:
Tibet, with its extreme altitude, is probably far colder than most regions much further north.
I reply:
As I said in an earlier comment: “one climate maps seems to show that its temperature is that of the southern taiga and the Alps, while another shows it not even as cold as Mongolia”.

Terry continued:
You are ignoring the apparent fact that Tibet was settled relatively recently, l
I reply:
The later it was settled the less likely anyone living there was there long enough to develop a new phenotype, so that point counts against it as the origin of the Mongoloid phenotype.

German Dziebel said...

@Tobus

"In which case we'd expect UI to have greater affinity to the ancestors of ancient East Asians and West Eurasians than to the ancestors of your postulated ancient Amerindians, no?"

And that's what you see on Fig. 2b in Fu et al.

"The bottlenecks postulated in Out Of Africa are all correlated with significant drops in heterozygosity. They are bottlenecks suggested *by* the data, not created out of thin air to cover holes in it."

Sure, pseudoscientist Tobus: the data shows drops in heterozygosity bit it doesn't show lifts in heterozygosity.

" and those means show that both both UI and Han are closer to Neanderthals than Amerindians are"

No, the Amerindians of 21th century follow UI (45,000 YBP) is having the longest Neandertal chunks. Just read my post, will you!

"The closest Sapiens to N/D is Loschbour, and Amerindians are closer to every other Sapiens than they are to N/D, so there's no "cline", just a big gap for you to fill with your imagination."

Like I said, you just don't believe in evolution. In your creationist mind, Africans were created by God and then human populations became increasingly Neandertal through bottlenecks as they migrated to America. Loschbour is an ancient sample. An ancient Amerindian sample will be more homozygous than Loschbour. In fact MA-1 was shown to be more homozygous than Loschbour, which is consistent with its Amerindian affinity. (https://verenich.wordpress.com/2014/11/09/%D0%BE%D1%86%D0%B5%D0%BD%D0%BA%D0%B0-%D0%B2%D0%BB%D0%B8%D1%8F%D0%BD%D0%B8%D1%8F-%D1%83%D1%80%D0%BE%D0%B2%D0%BD%D1%8F-%D0%B0%D1%83%D1%82%D0%BE%D1%81%D0%BE%D0%BC%D0%BD%D0%BE%D0%B9-%D0%B3%D0%BE%D0%BC/).

http://www.nationsonline.org/oneworld/languages.htm

Read some science: http://www.mapageweb.umontreal.ca/tuitekj/cours/IE/annurev.anthro.NICHOLS.pdf.

"That "long time ago" is only a fraction of the period of time in which they accumulated heterozygosity as a single species - it's at most 25% of the time since homo divergence from chimps (using very generous 4mil chimp/homo and 1mil neand/sap dates, more like 6% if we use the 8mil/0.5mil extremes). The "add up" will be heavily skewed towards lost heterozygosity and the "minimal" increase would in fact be negative - a net *decrease* in heterozygosity."

Sheer nonsense. You look smarter when you have pebbles in your hands.

"That's the important difference between your theory and OOA, and why your theory is fringe and crackpot while OOA is universally accepted."

Speciation occurs on the fringes. And OOA is "universally accepted" only because geneticists forced their data into a crackskulls' framework.

terryt said...

"you cannot deny that the movement from SW Asia to S Asia is simpler than movement from SW Asia through S Asia to SE Asia and then back again to SC Asia".

Simpler, but by no means less likely when you consider ALL the evidence. Under your proposed scenario at least ten branches of Y-DNA K2 have to move, fully-formed, from SW Asia all the way to New Guinea. To me that hypothesis is extremely unlikely.

"That K2 is primarily is a SE/E Asian bunch of haplogroups is what you must prove"

Use your own eyes, not just make up stuff. K2b2 is the ONLY K2 haplogroup found outside SE Asia and nearby regions. Have you actually seen the update of the K clade? If not I advise you look it up.

"still leaves most of China within the Palearctic: all of north China and some of south China".

I presume you mean this map:

http://en.wikipedia.org/wiki/Ecozone

Check your atlas. According to that map all of Europe, the Middle East and even North Africa are part of the Palearctic zone. Are you claiming these regions have the same climate as Central Asia? And basically the only part of China in the Palearctic zone is Tibet and Inner Mongolia, plus a little of the North China Plain. I totally agree that those regions are Palearctic, but they unlikely to be as far north as where NO originated anyway.

"I will point that it that is all the more reason for thinking that people moved from Central Asia to East Asia rather than the reverse, since they would be moving downhill rather than uphill!"

What? I assume you're joking. You have to get up the hill before you can come down. Do you not know the rhyme about the grand old Duke of York?

"So we now have the branches of K2a as follows:
Ust-Ishim—found near the border of C Asia and N Asian
NOX
X—found only in Indi
NO
N—found mainly in N Asia, and not found south of E Asian
O—the only one found in SE Asia"

As Ebizur says, Ust-Ishim belongs to a branch of K2a that is outside NOX, but it is still within K2a, not a separate branch of K2.

"lest you claim that my grouping SW, C, N and E Asia together against SE Asia is arbitrary, the first four share the property of fitting comfortably with my theory, whereas only SE Asia does not".

And all part of just one clade within K2a. So we have one clade (K2a) spread from northeast India, through China and into Central Asia. All other K2 clades except K2b2 are confined to SE Asia and New Guinea/Australia. So your four haplogroups do not share the property of fitting comfortably with your theory at all. You are refusing to consider any other K2 haplogroup, presumably because to do so would immediately invalidate your theory. The only two Central Asian K2 haplogroups are outliers from two separate K2 clades. The two are not even closely related, a fact you are intent on ignoring.

Tobus said...

@German:
And that's what you see on Fig. 2b in Fu et al.

Check out Table S11.2

the data shows drops in heterozygosity bit it doesn't show lifts in heterozygosity

The point is that these drops actually exist, the ones you made up don't.

the Amerindians of 21th century follow UI (45,000 YBP) is having the longest Neandertal chunks.

We were talking about genetic affinity, not length of chunks.

Like I said, you just don't believe in evolution. In your creationist mind, Africans were created by God

Ha! You're an idiot.

Read some science:

I quote: "These principles apply equally well in areas of earliest human inhabitation (Africa), continents colonized early(Australia), continents colonized recently (the Americas)".. let me guess, not science anymore?

Sheer nonsense.

Are you saying sapiens and N/D *haven't* shared the majority of their evolution?!?

And OOA is "universally accepted" only because geneticists forced their data into a crackskulls' framework

I disagree, it's universally accepted because it's the best fit for both the genetic *and* archaeological data. Both fields lead us independently to the same conclusion, there's no forcing needed.

Gregory76 said...

Terry said:
Under your proposed scenario at least ten branches of Y-DNA K2 have to move, fully-formed, from SW Asia all the way to New Guinea.

I reply:
I do not propose 10 branches of K2 moved to New Guinea. I only propose that 2 branches, K2c and K2d, and one subbranch, K2b1 (PMS) moved there.


I said earlier:
"That K2 is primarily is a SE/E Asian bunch of haplogroups is what you must prove"

Terry said:
Use your own eyes, not just make up stuff.

I reply:
I just told you that you must prove a claim, so the question is whether you are making stuff up to support that claim or not, not whether I am making up stuff or not, since I did not advance a new claim nor, in that clause, repeat an old one.

Terry said:
K2b2 is the ONLY K2 haplogroup found outside SE Asia and nearby regions.

I reply:
Clearly not, since N is found in N Asia and C Asia, and Ust-Ishim in southwest Siberia.


Terry continued:
And basically the only part of China in the Palearctic zone is Tibet and Inner Mongolia, plus a little of the North China Plain.

I reply:
Certainly not. Check that map again. The whole of North China plain and part of the South China plain. And on Wallace's map it was a majority of South China plain that was included in the Palearctic.
Also, if consider the whole of East Asia and not just China, Korea and Japan are Palearctic, too.


I said earlier:
"I will point that it that is all the more reason for thinking that people moved from Central Asia to East Asia rather than the reverse, since they would be moving downhill rather than uphill!"

Terry replied:
What? I assume you're joking. You have to get up the hill before you can come down. Do you not know the rhyme about the grand old Duke of York?

I reply:
I must assume you're joking. I'm not familiar with rhyme, but it does not change the fact that it would easier from people from Central Asia to move down to East Asia than the reverse.
Yes, ultimately people must have reached the highest altitudes from lower ones, but it would be easier from people from SW Asia, which has some highland areas, to move to C Asia, and then move down into the China plain, than for people from the China plain to move into the highland areas of C Asia.


Terry said:
As Ebizur says, Ust-Ishim belongs to a branch of K2a that is outside NOX, but it is still within K2a, not a separate branch of K2.

I reply:
Ust-Ishim being outside of NOX, rather than being N, NO or even NOX, means that it is not merely one of four branches of K2a, nor one of three, but rather one of two. So Ust-Ishim gives us at least as much reason to claim that K2a originated far from SE Asia as not, even if the case for believing NOX originated in SE Asia is very strong, which it is not.

I said earlier:
"lest you claim that my grouping SW, C, N and E Asia together against SE Asia is arbitrary, the first four share the property of fitting comfortably with my theory, whereas only SE Asia does not".

Terry replied:
And all part of just one clade within K2a.

I reply:
No, those are all regions of Asia, not clades. My point is that all of those regions share the important characteristic of being locations for haplotypes that are easily compatible with my theory.

Terry continued:
So we have one clade (K2a) spread from northeast India, through China and into Central Asia.

I reply:
Again, that is claim that must be proved, and it will be hard. One primary branch has been found only in southwest Siberia.

Terry continued:
All other K2 clades except K2b2 are confined to SE Asia and New Guinea/Australia.

I reply:
That sounds like a lot, until you realize “All other” means just K2c and K2d, and the one branch of K2b, and perhaps the two unresolved cases.

Terry continued:
The only two Central Asian K2 haplogroups are outliers from two separate K2 clades. The two are not even closely related, a fact you are intent on ignoring.

I reply:
Their being from two separate clades and not being closely related counts for my theory, not yours.

terryt said...

"I do not propose 10 branches of K2 moved to New Guinea. I only propose that 2 branches, K2c and K2d, and one subbranch, K2b1 (PMS) moved there".

That in itself seems extremely unlikely. The idea that the three branches should have moved all the way from SW Asia and then sorted themselves out as to where in SE Asia they lived is impossible to accept. Who did the drafting? God? And once more you're conveniently ignoring K2* in SE Asia. And I also assumed you would believe K2b1a, K2b1b, K2b1c and M as well as the various branches of K2a would have been fully-formed in SW Asia. Why do you not believe that to be so?

"That sounds like a lot, [All other K2 clades except K2b2 are confined to SE Asia and New Guinea/Australia] until you realize 'All other' means just K2c and K2d, and the one branch of K2b, and perhaps the two unresolved cases".

Well, that's five clades. Meanwhile we just have two subclades of two K2 branches found outside that region.

"Clearly not, since N is found in N Asia and C Asia, and Ust-Ishim in southwest Siberia".

You're manipulating the facts again. Neither of those haplogroups is a basal branch of K2. They are both branches within K2a, with N even being a sub-branch of NO.

"Ust-Ishim being outside of NOX"

Yes. Outside NOX but still within K2a.

"Their being from two separate clades and not being closely related counts for my theory, not yours".

It actually totally destroys your theory. Their separate identity indicates two separate movements into Central Asia, not a single original one that lies at the ancestry of all the K2 clades.

"I just told you that you must prove a claim"

I have proved it but you refuse to see. There are none so blind as those who will not see.

"ultimately people must have reached the highest altitudes from lower ones, but it would be easier from people from SW Asia, which has some highland areas, to move to C Asia, and then move down into the China plain, than for people from the China plain to move into the highland areas of C Asia".

Is that right? I can only assume you are again joking.

terryt said...

Just to spell it out. I presume you are prepared to agree there are five basal K2 haplogroups: K2*, K2a, K2b, K2c and K2d. Three of these basal groups (K2*, K2c and K2d) are confined to Southeast Asian islands. That's a majority for a start, three out of five.

K2b divides into two, one branch of which is also confined to Southeast Asian islands. East of Wallace's Line in fact. The other branch is Central Asian but that alone does not indicate in any way at all that the split between the two occurred there. That's three and a half out of five in Southeast Asia and one half in Central Asia. An even bigger majority for Southeast Asia.

Turning to K2a. As you say it too divides into two: NOX and Ust'-Ishim K2a*. That is another half haplogroup in Central Asia, giving the equivalent of one haplogroup out of five in Central Asia. NOX also divides into two: NO and X. X is northeast India, just to the northwest of Southeast Asia. So that's another quarter haplogroup for Southeast Asia. NO also divides into two: N and O. N is Central Asia and so that adds one eighth of a haplogroup for Central Asia and one eighth for China, just to the north of Southeast Asia.

Therefore we finish up with one and one eighth haplogroups for Central Asia and three and seven eighths for Southeast Asia and immediate environs. A clear majority for Southeast Asia.

German Dziebel said...

@Tobus

"The point is that these drops actually exist, the ones you made up don't."

Quite the opposite. You made your drops up as part of your biblical story of descent from the Garden of Eden. The hikes in fact are supported by ancient DNA data.

"We were talking about genetic affinity, not length of chunks."

That's affinity, my creationist friend! That's how it was established that Ust-Ishim is Neandertal admixed. The Amerindian connection is abundantly supported by other data.

"I disagree, it's universally accepted because it's the best fit for both the genetic *and* archaeological data. Both fields lead us independently to the same conclusion, there's no forcing needed."

They didn't come to it independently. You think geneticists just walked into human prehistory without consulting with archaeologists and paleobiologists? Typical creationist ignorance of how science works.

"Ha! You're an idiot."

No, I'm a professional with two doctorates. You are a Gospel-reading freak. There are a bunch of those out there.

German Dziebel said...

@Tobus

"The point is that these drops actually exist, the ones you made up don't."

Quite the opposite. You made your drops up as part of your biblical story of descent from the Garden of Eden. The hikes in fact are supported by ancient DNA data.

"We were talking about genetic affinity, not length of chunks."

That's affinity, my creationist friend! That's how it was established that Ust-Ishim is Neandertal admixed. The Amerindian connection is abundantly supported by other data.

"I disagree, it's universally accepted because it's the best fit for both the genetic *and* archaeological data. Both fields lead us independently to the same conclusion, there's no forcing needed."

They didn't come to it independently. You think geneticists just walked into human prehistory without consulting with archaeologists and paleobiologists? Typical creationist ignorance of how science works.

"Ha! You're an idiot."

No, I'm a professional with two doctorates. You are a Gospel-reading freak. There are a bunch of those out there.

Tobus said...

@German:
Quite the opposite.

Is there empiricial data that shows a "hike" between K14 and MA-1... or did you just make it up?

That's how it was established that Ust-Ishim is Neandertal admixed.

No, "To determine whether the Ust’-Ishim individual has archaic admixture we used the D statistic" (SI16). It's these D-stats that show CHB is the "most Neanderthal" of modern pops.

They didn't come to it independently

Who's "they"? I said the fields both lead us there independently - archaeology places the earliest modern humans in Africa, genetics places the MCRA of modern humans in Africa. Both of these conclusions are drawn independently from different sets of data, neither is "forced" into that position by the other.

No, I'm a professional with two doctorates. You are a Gospel-reading freak.

If you honestly think I'm a "Gospel-reading freak" then you are an idiot, two doctorates or twenty.

Gregory76 said...

Terry,
You are again counting East Asia and South Asia with Southeast Asia: SE Asia counts against my theory, but E Asia and S Asia are compatible with my theory.
Your theory has problems in addition to its postulating unnecessarily long migration routes, its difficulty in accounting for the formation of the Mongoloid phenotype, and its leading you into misreading the zoogeographic maps and to arguing in a circle.
It posulates a northern migration from Southeast Asia to East Asia when all migrations known from history and archeology have been from East Asia to Southeast Asia. Tai peoples moved south from China a few centuries ago, Austronesians spread from Taiwan to Southeast Asia, Tibeto-Burmans moved south from Tibet and SE China into India and Burma, and I think Austroasiatics also moved from E Asia southward, though might disagree on that. I find it odd that, while all of those groups saw fit to move south, your theory postulates an additional, earlier migration north. Now you might try to say that it was not 4 (or 3) independent migrations: that the Chinese pushed these people south. But you cannot say that about the Tibeto-Burmans, nor about the Austroasiatics, if you say that they were never in E Asia. And while the Tai and Austronesian may have been one group broken apart by one southern migration of the Chinese, with the Tai in the east and Austronesian heading to Taiwan, you can hardly say that the migration from Taiwan was the direct result of Chinese pressure. So you must postulate that the ancestors of the Tai and the Austronesian, as one group in one migration or two groups in two migrations, moved north from SE Asia into China, and their descendents later migrated in the reverse direction. And if you admit that Austroasiatics moved south from China at some time then you must postulate two migrations for them, too. Indeed, if suppose that the mass of the Sino-Tibetans, being O, also moved north from SE Asia, then you have 2 migrations for them, since the Tibeto-Burmans and Chinese both moved south. This theory is implausibly complex.

terryt said...

"You are again counting East Asia and South Asia with Southeast Asia: SE Asia counts against my theory, but E Asia and S Asia are compatible with my theory".

Your imagination is running wild here. Do you believe there is some impassable barrier between SE Asia and the two neighboring regions? East and South Asia share a single branch of K2 in the form of K2a. NO in East Asia and X in South Asia. Surely we should expect some leakage from SE Asia.

"Your theory has problems in addition to its postulating unnecessarily long migration routes"

Which 'unnecessarily long migration routes'. K is sufficiently later than CT to have moved almost everywhere accessible by the time K2 first formed.

"its difficulty in accounting for the formation of the Mongoloid phenotype"

The development of the Mongoloid phenotype is an independent development and confined to somewhere near northern China. It is certainly not confined to a single Y-DNA either.

"all migrations known from history and archeology have been from East Asia to Southeast Asia".

Stated with such confidence. It is obvious that several South Asian haplogroups entered Burma and Thailand during prehistory and the Australian Aborigine phenotype has often been compared to that of South Asian tribals. Both groups lack the Mongoloid phenotype and so any movement from Australia to South Asia must predate its development. Prehistory was far more complicated than a simple human expansion like ink through a blotter.

"Tai peoples moved south from China a few centuries ago, Austronesians spread from Taiwan to Southeast Asia, Tibeto-Burmans moved south from Tibet and SE China into India and Burma, and I think Austroasiatics also moved from E Asia southward"

Those migrations are all post Paleolithic, probably within the last 10,000 years.

terryt said...

"I find it odd that, while all of those groups saw fit to move south, your theory postulates an additional, earlier migration north".

So you believe all human migrations have always been in the same direction?

"Now you might try to say that it was not 4 (or 3) independent migrations: that the Chinese pushed these people south. But you cannot say that about the Tibeto-Burmans"

Why not? The Tibeto-Burman language is basically a branch of Sino-Tibetan. Speakers of Tibeto-Burman also display a level of Mongoloid phenotype and so presumably came from the north somewhere.

"So you must postulate that the ancestors of the Tai and the Austronesian, as one group in one migration or two groups in two migrations, moved north from SE Asia into China"

I postulate no such thing. They did not move 'north' from anywhere. However their 'ancestors', in the form of K2a, had much earlier moved north from SE Asia. Tai and Austronesian are a later movement of two closely related languages, probably an early Neolithic movement originally associated with Y-DNA O1. During the Early Neolithic O1 formed a majority of Y-DNA around the mouth of the Yantze and so quite probably expanded south into SE Asia from there.

"you can hardly say that the migration from Taiwan was the direct result of Chinese pressure".

Agreed. It was the direct result of improved boating and trading around the South China Sea which ultimately allowed movement direct from Taiwan to the Philippines. And rapidly beyond.

"And if you admit that Austroasiatics moved south from China at some time then you must postulate two migrations for them, too".

Why? The two groups are separated by the fact one moved by land (Tai) and the other by sea (Austronesian).

"if suppose that the mass of the Sino-Tibetans, being O, also moved north from SE Asia, then you have 2 migrations for them, since the Tibeto-Burmans and Chinese both moved south. This theory is implausibly complex".

'Complex' because what you believe is preventing you from seeing the simplicity. The 'southward' movement involved derived haplogroups. The 'original' movement north had involved the basal K2a. I've asked before but on what grounds do you believe that human migration has always been unidirectional?

Gregory76 said...

I said earlier: "You are again counting East Asia and South Asia with Southeast Asia: SE Asia counts against my theory, but E Asia and S Asia are compatible with my theory".
Terry replied: Do you believe there is some impassable barrier between SE Asia and the two neighboring regions?
I reply: No. My theory is that K2 originated in and began diversifying in SW Asia or C Asia, and reached SE Asia only trough C Asia and E Asia. So early haplotypes in SE Asia count against it, but those in E Asia or S Asia easily fit with my theory

I said earlier: "Your theory has problems in addition to its postulating unnecessarily long migration routes"
Terry replied: Which 'unnecessarily long migration routes'.
I reply: R reaching western Eurasia, N reaching N Asia, and X reaching S Asia, all by way of SE Asia, when they could be all explained by shorter routes from SW Asia.

I said earlier: "its difficulty in accounting for the formation of the Mongoloid phenotype"
Terry replied: The development of the Mongoloid phenotype is an independent development and confined to somewhere near northern China.
I reply: Again, read my earlier objections regarding Tibet (let alone N China or S China) not being cold enough, and the current inhabitants not being as Mongoloid in their features as more northern areas, which would require some earlier group to be the first Mongoloids, yet yourself pointed out that Tibet was only settled relatively recently.

I said earlier: "all migrations known from history and archeology have been from East Asia to Southeast Asia".
Terry replied:
It is obvious that several South Asian haplogroups entered Burma and Thailand during prehistory and the Australian Aborigine phenotype has often been compared to that of South Asian tribals.
I reply: I was talking about the later, Mongoloid migrations, involving descendents of K2

I said earlier:
"I find it odd that, while all of those groups saw fit to move south, your theory postulates an additional, earlier migration north".
Terry replied: So you believe all human migrations have always been in the same direction?
I reply: No, but I believe we should not postulate an unnecessary amount migrations, not only from consideration not only of Ockham's Razor but also of the fact that back then survival during migration was arduous.

I said earlier: "So you must postulate that the ancestors of the Tai and the Austronesian...moved north from SE Asia into China"
Terry replied: I postulate no such thing. They did not move 'north' from anywhere. However their 'ancestors', in the form of K2a, had much earlier moved north from SE Asia.
I reply: That's just what I said: the ancestors of those two groups from north from SE Asia. Read more carefully.

I said earlier:"And if you admit that Austroasiatics moved south from China at some time then you must postulate two migrations for them, too".
Terry replied: Why? The two groups are separated by the fact one moved by land (Tai) and the other by sea (Austronesian).
I reply: Read more carefully: I said “Austroasiatic”, nor “Austronesian”.

I said earlier: "This theory is implausibly complex".
Terry replied:
….The 'southward' movement involved derived haplogroups. The 'original' movement north had involved the basal K2a.
I reply:
That is still two migrations in that lineage. It doesn't matter how many haplotype mutations occurred on the route from Africa to the current residence of a given lineage: it is still the same path. And a shorter path, all else being equal, is more likely.

terryt said...

"And a shorter path, all else being equal, is more likely".

Only if you're committed to and fixated on a Noah's Ark perspective on human evolution and expansion. Everything moving in one direction from a single place.

"I believe we should not postulate an unnecessary amount migrations, not only from consideration not only of Ockham's Razor but also of the fact that back then survival during migration was arduous".

We should not postulate 'no migration' when phylogenies indicate otherwise. And I think we can assume that migrations occurred when conditions were favourable. It was only during periods of climate change or extinction of resources that survival became 'arduous'. Multiple haplogroups, both mt- and Y-, indicate a period of expansion followed by a period of consolidation far from where the haplotype almost certainly originated.

"My theory is that K2 originated in and began diversifying in SW Asia or C Asia, and reached SE Asia only trough C Asia and E Asia".

I see no evidence at all for such a theory. You seem to be completely ignoring the actual phylogeny of K. The only K2 haplogroup anywhere near SW Asia is the very derived R, a member of K2b. It is extremely unlikely K2b itself originated in SW Asia, although it's quite possible R did so.

"So early haplotypes in SE Asia count against it, but those in E Asia or S Asia easily fit with my theory"

Except there are no 'early' K2 haplotypes anywhere other than SE Asia apart from the widespread K2a, which is found very near SE Asia. I've tried to point out on numerous occasions the only K2 haplogroups outside SE Asia are derived ones but you are unable to see that for some reason.

"R reaching western Eurasia, N reaching N Asia, and X reaching S Asia, all by way of SE Asia, when they could be all explained by shorter routes from SW Asia".

Except that explanation doesn't fit the phylogeny. And of those haplogroups only X could be said to have moved directly from SE Asia anyway, involving a very short jump into the far northeastern region of South Asia. Some basal P is evidently found in SE Asia after all, and it is probable QR's ancestor moved through South Asia, quite close to SE Asia, surprisingly. But R itself certainly did not originate in SE Asia. Neither did N. N moved into northern Asia from East Asia where NO had become well established long before N or O even developed.

terryt said...

"which would require some earlier group to be the first Mongoloids"

My bet is that the Mongoloid phenotype introgressed into the modern humans from an ancient population somewhere near the Tibet/Mongolia/China border region. In any case that is where it seems to have first entered the modern human gene pool. And we don't know that cold alone is responsible for the phenotype anyway. My guess is it is the result of a highly reflective environment. After all semi-desert dwelling Khoisan have a fairly distinct eye-fold.

"the current inhabitants not being as Mongoloid in their features as more northern areas"

Almost certainly because the Mongoloid phenotype is mixed with a pre-Mongoloid phenotype in southern China.

"I was talking about the later, Mongoloid migrations, involving descendents of K2"

The vast majority of populations carrying K2 haplogroups are not Mongoloid in any way. Such are confined to the downstream haplogroups Q, N, O1, O2, and O3. These obviously picked up, or developed, the Mongoloid phenotype after their ancestors had moved north. And the non-K2 haplogroup C2 is closely associated with the phenotype. I agree O1, O2, O3 and C2 are all later southward moving haplogroups.

"That's just what I said: the ancestors of those two groups from north from SE Asia. Read more carefully".

You should read more carefully yourself. I said the ancestors of those groups moved north from near SE Asia, not the groups themselves. The haplogroups O1, O2 and O3 were not the haplogroups that moved north. They are the ones that moved back south (probably with the expansion of the Chinese Neolithic) after developing somewhere in northern or central China, from the K2a that had much earlier moved north from near SE Asia.

"Read more carefully: I said 'Austroasiatic', nor 'Austronesian'".

Sorry. I did mistake that one. However Austroasiatic does appear to have moved south over land from the central Yangtze River region. Austro-Asiatic is closely associated with Y-DNA O2, Austronesian with O1 and Sino-Tibetan with O3. Those connections would not exist if the movement was at all as ancient as you are trying to suggest.

"That is still two migrations in that lineage".

It is actually only one migration for each lineage. K2a moved north, where NO developed somewhere in the region that the Chinese Neolithic later developed. Then O1, O2 and O3 moved south independently of each other and N moved north. I would have thought it was simplicity itself.

terryt said...

Sorry. Correction:

"Sino-Tibetan with O3"

Actually Sino-Tibetan is associated specifically with O3a2c. The related O3a2b is associated with the Hmong-Mien. But Hmong-Mien looks to have accumulated O2a somewhere along the line. Almost certainly through the O3a2b population's mixing with earlier inhabitants as the language and genes expanded. In other words O2a's expansion was earlier than that of O3a2b's.

By the way, under your scenario there are no 'earlier' inhabitants in SE Asia for your postulated southward moving haplogroups to mix with. Have you not considered that problem?

Gregory76 said...

I said earlier:
“And a shorter path, all else being equal, is more likely".
Terry replied:
Only if you're committed to and fixated on a Noah's Ark perspective on human evolution and expansion. Everything moving in one direction from a single place.
I reply:
No, whether I am commited to such a view or not, and I am not (as I said earlier), it is still the case that the shorter path, all else being equal, is more likely. The shorter path is simpler, and the simpler theory, all else being equal, is more likely to be true, by Ockham's Razor. Furthermore, the long path requires more effort, time and resources, and the effort was considerable.

I said earlier:
"I believe we should not postulate an unnecessary amount migrations, not only from consideration not only of Ockham's Razor but also of the fact that back then survival during migration was arduous".
Terry replied:
We should not postulate 'no migration' when phylogenies indicate otherwise.
I reply:
Phylogenies alone do not. Phylogenies plus the assumptions of your theory about where some branches on the tree lived at a given time may indicate additional migrations, but that theory needs to be proven, not assumed in the process of trying to prove it.
Terry continued:
And I think we can assume that migrations occurred when conditions were favourable. It was only during periods of climate change or extinction of resources that survival became 'arduous'.
I reply:
Certainly not. Migration, and most aspects of survival, were arduous for most people, down to at least the Industrial Revolution.
Terry continued:
Multiple haplogroups, both mt- and Y-, indicate a period of expansion followed by a period of consolidation far from where the haplotype almost certainly originated.
I reply:
Yes, and sometimes it was during the consolidation in the new home that many of the younger haplotypes originated.

I said earlier:
"My theory is that K2 originated in and began diversifying in SW Asia or C Asia, and reached SE Asia only trough C Asia and E Asia".
Terry replied:
I see no evidence at all for such a theory.
I reply:
I have given evidence several times, including the shorter migration paths in my theory, the minimizing of retrograde motion, its better fitting with the correction explanation of the original of the Mongoloid phenotype, and considerations of zoogeography.

Terry continued:
The only K2 haplogroup anywhere near SW Asia is the very derived R.
I reply:
You have to admit, and did admit earlier, that haplotypes came out of Africa by way of SW Asia.

I said earlier:
"So early haplotypes in SE Asia count against it, but those in E Asia or S Asia easily fit with my theory"
Terry replied:
Except there are no 'early' K2 haplotypes anywhere other than SE Asia apart from the widespread K2a, which is found very near SE Asia.
I reply:
You have interpreted 'early' too narrowly. I was counting all of the K2 haplotypes under discussion as early, except O in SE Asia. So K2c and K2d count against my theory, but not the others, because they are present in places that can be accomodated by my theory.

Terry continued:
I've tried to point out on numerous occasions the only K2 haplogroups outside SE Asia are derived ones but you are unable to see that for some reason.
I reply:
All of the y haplotypes except A are 'derived' from other haplotypes. 'Derived' is a matter of degree, as is 'early'.

Gregory76 said...

said earlier: "R reaching western Eurasia, N reaching N Asia, and X reaching S Asia, all by way of SE Asia, when they could be all explained by shorter routes from SW Asia".
Terry replied: Except that explanation doesn't fit the phylogeny.
I reply: It does. See my above remark about phylogeny v. phylogeny+assumptions.

Terry continued: And of those haplogroups only X could be said to have moved directly from SE Asia...into the far northeastern region of South Asia.
I reply: No matter how short the migration from S Asia to SE Asia, the migration from SW Asia to S Asia will be shorter than a one from SW Asia through S Asia to SE Asia and back to S Asia.

Terry continued: Some basal P is evidently found in SE Asia after all, and it is probable QR's ancestor moved through South Asia, quite close to SE Asia, surprisingly.
I reply: You assume that basal lineages never move with their descendents.

I said earlier: "which would require some earlier group to be the first Mongoloids"
Terry replied: My bet is that the Mongoloid phenotype introgressed into the modern humans from an ancient population somewhere near the Tibet/Mongolia/China border region.
I reply: Saying that it originated with a pre-AMH population does solve the problem.
Terry continues: In any case that is where it seems to have first entered the modern human gene pool.
I reply: What evidence do you have for that? As I said before, the area does not seem to be cold enough.

Terry continues: And we don't know that cold alone is responsible for the phenotype anyway. My guess is it is the result of a highly reflective environment. After all semi-desert dwelling Khoisan have a fairly distinct eye-fold.
I reply: That may work for the eyefold (although I think high winds is a more likely explanation), but not most other Mongoloid features, since the Khoisan phenotype is far from Mongoloid.

I said earlier: "the current inhabitants not being as Mongoloid in their features as more northern areas"
Terry replied: Almost certainly because the Mongoloid phenotype is mixed with a pre-Mongoloid phenotype in southern China.
I reply: If the Mongoloid phenotype originated in Tibet, then how does its dilution by later mixture in S China account for the Tibetans being less Mongoloid than northern Mongoloids?

I said earlier: "That's just what I said: the ancestors of those two groups from north from SE Asia. Read more carefully".
Terry replied: You should read more carefully yourself. I said the ancestors of those groups moved north from near SE Asia, not the groups themselves.
I reply: I did read carefully and you did not: I said “So you must postulate that the ancestors of the Tai and the Austronesian,...moved north from SE Asia into China”, and you replied “I postulate no such thing....However their 'ancestors', in the form of K2a, had much earlier moved north from SE Asia.”, and that is what I said: the ancestors of those 2 groups moved north.

Terry said: Austro-Asiatic is closely associated with Y-DNA O2, Austronesian with O1 and Sino-Tibetan with O3. Those connections would not exist if the movement was at all as ancient as you are trying to suggest.
I reply: I am not supposing that the movement was as old as the alleged movement from SE Asia to E Asia is supposed by your theory to be.

I said earlier: "That is still two migrations in that lineage".
Terry replied: It is actually only one migration for each lineage.
I reply: I said “lineage”, not “haplotype”. By 'lineage', I mean a line of descent, such as could be traced by any human to the first human. Every time a new haplotype mutation occurs in the lineage, we have a new haplotype, but still the same lineage. So you postulate the lineage running from the earliest K2a to the modern inhabitants of SE Asia as migrating from SE Asia to E Asia and later migrating back to SE Asia, whereas I postulate a migration from N or C Asia through E Asia to SE Asia without a reversal of direction.

terryt said...

"All of the y haplotypes except A are 'derived' from other haplotypes. 'Derived' is a matter of degree, as is 'early'"

Well of course you are free to call R1b1a2a1a1 'early' if you insist but it obviously arose considerably later than A.

"You have interpreted 'early' too narrowly. I was counting all of the K2 haplotypes under discussion as early, except O in SE Asia".

On what grounds did you not consider O? After all you seem to include N. And R is certainly no 'earlier' than either of those haplogroups. Your counting system is wrong anyway. There are actually only four 'early' K2 Y-DNAs: K2a, K2b, K2c and K2d. All the other ones you insist on including are derived from one or other of these, and therefore can only be 'later'. It is an open question yet as to how much later of course but two are confined to what is now island SE Asia and the other two have connection there or very nearby. Both X and Ust'-Ishm K, as well as N and O, derive from K2a and Q'R derives from K2b.

"You have to admit, and did admit earlier, that haplotypes came out of Africa by way of SW Asia".

Everyone agrees on that. But migration after leaving Africa would hardly have been a simple process with humans filling up the whole of the earth. They would have moved through the most desirable regions first, which may have taken them long distances along narrow migration routes. We know they reached SE Asia and even Australia reasonably soon after leaving Africa (given a date for that leaving of 55,000 years that many accept). That's a pretty long leap and proves such leaps were far from impossible 'early' on.

"Certainly not. Migration, and most aspects of survival, were arduous for most people"

Yes. That's why conditions had to be ideal for them to go ahead.

"The shorter path is simpler, and the simpler theory, all else being equal, is more likely to be true, by Ockham's Razor".

The shorter path is not necessarily the easiest though.

"the shorter migration paths in my theory, the minimizing of retrograde motion, its better fitting with the correction explanation of the original of the Mongoloid phenotype, and considerations of zoogeography".

A route through Central Asia is hardly 'shorter' that one through South Asia. And the Mongoloid phenotype evolved some time after humans had reached East Asia and so says nothing about a route to that region. And why do you find it necessary to minimise retrograde motion? We know humans have moved backwards and forwards through Central Asia and South Asia many times.

"the long path requires more effort, time and resources, and the effort was considerable".

Quite. Although in fact I think Y-DNA C took a Central Asian route east. And probably arrived in Australia before K did. But that's another subject.

terryt said...

I think we're wasting time here if you're going to just refuse to consider phylogeny.

"See my above remark about phylogeny v. phylogeny+assumptions".

The huge assumption you are making is that K2c and K2d both originated in SW Asia, and then migrated all the way through Central Asia to take up separate residency in two separate islands in SE Asia. That is so unlikely as to be not even worth consideration. You are only able to make that assumption by completely ignoring the basal position ofthose two haplogroups.

"I am not supposing that the movement was as old as the alleged movement from SE Asia to E Asia is supposed by your theory to be".

But your theory makes it impossible for anyone to have been in SE Asia before the O haplogroups moved south.

"What evidence do you have for that? As I said before, the area does not seem to be cold enough".

Another assumption on your part. You assume 'cold' is the explanation for the Mongoloid phenotype. The evidence for its region of origin, or at least a substantial portion of the phenotype is here:

http://blogs.discovermagazine.com/gnxp/2013/02/is-girls-generation-the-outcome-of-the-pleistocene-mind/

Check out diagram B.

"If the Mongoloid phenotype originated in Tibet, then how does its dilution by later mixture in S China account for the Tibetans being less Mongoloid than northern Mongoloids?"

I didn't say it originated 'in' Tibet. I said (and you actually quoted it) near the 'Tibet/Mongolia/China border region'. We know from phylogenies (note) that much of the Tibetan gene pool entered that country from the southeast. You obviously know very little indeed about East and Se Asian prehistory.

"So you postulate the lineage running from the earliest K2a to the modern inhabitants of SE Asia as migrating from SE Asia to E Asia and later migrating back to SE Asia"

Yes. What do you see as the problem?

"whereas I postulate a migration from N or C Asia through E Asia to SE Asia without a reversal of direction".

Which does not fit archaeology, anthropology or genetics. You are just making it up.

Gregory76 said...

I said: "You have to admit, and did admit earlier, that haplotypes came out of Africa by way of SW Asia".
Terry replied: Everyone agrees on that. But....[humans] would have moved through the most desirable regions first, which may have taken them long distances along narrow migration routes.
I reply: Yes. Are you claiming that a route through S Asia and SE Asia goes through more desirable regions than one through C Asia?

I said: "Certainly not. Migration, and most aspects of survival, were arduous for most people"
Terry replied: Yes. That's why conditions had to be ideal for them to go ahead.
I reply: And ideal conditions are rare.

Terry said: The shorter path is not necessarily the easiest though.
I reply: What makes you think the path in your theory is the easier one?

Terry said: A route through Central Asia is hardly 'shorter' that one through South Asia.
I reply: It is a shorter route to Europe and western Siberia, and, to a lesser extent, to eastern Siberia and E Asia.

Tery said: And the Mongoloid phenotype evolved some time after humans had reached East Asia and so says nothing about a route to that region.
I reply: It says that it had to go through a cold region.

Terry said: And why do you find it necessary to minimise retrograde motion?
I replied: Because it longer and because it is more arduous.

Terry said: The huge assumption you are making is that K2c and K2d both originated in SW Asia,
I reply: No. It is possible that they arose from K2 after K2 had migrated to SE Asia.

Terry said: ...your theory makes it impossible for anyone to have been in SE Asia before the O haplogroups moved south.
I said: Why would you think that?

Terry said: You assume 'cold' is the explanation for the Mongoloid phenotype.
I reply: Not an assumption. It is based on that fact that the closer a body shape is to a sphere the less it tends to lose heat. So cold regions select for short, heavyset bodies; short, heavy limbs, extremities and digits; round heads; round faces; high cheekbones; and small, concave noses, all of which characterize the Mongoloid phenotype.

Terry says: The evidence for its region of origin, or at least a substantial portion of the phenotype is here: http://blogs.discovermagazine.com/gnxp/2013/02/is-girls-generation-the-outcome-of-the-pleistocene-mind/ Check out diagram B.
I reply: EDAR is shown as uncommon in Siberia, where the Mongoloid type is strongest.

I said: "If the Mongoloid phenotype originated in Tibet, then how does its dilution by later mixture in S China account for the Tibetans being less Mongoloid than northern Mongoloids?"
Terry replied: I didn't say it originated 'in' Tibet. I said...near the 'Tibet/Mongolia/China border region'.
I reply: That is even less likely: China is warmer than Tibet.

I said: "So you postulate the lineage running from the earliest K2a to the modern inhabitants of SE Asia as migrating from SE Asia to E Asia and later migrating back to SE Asia"
Terry replied: Yes. What do you see as the problem?
I reply: Too much distance and too much effort required.

I reply: "whereas I postulate a migration from N or C Asia through E Asia to SE Asia without a reversal of direction".
Terry replied: Which does not fit archaeology, anthropology or genetics.
I reply: It is what most physical anthropologists would have said before the deliverances of modern genetics in the last three decades, and those deliverances do not compel its abandonment.

terryt said...

"Are you claiming that a route through S Asia and SE Asia goes through more desirable regions than one through C Asia?"

Certainly warmer.

"And ideal conditions are rare".

Especially so through Central Asia, although I do agree it has often provided a route. The route through South Asia certainly would not have included the whole subcontinent. Jungle, desert and the Ganges/Brahmaputra Delta would have restricted the available route.

"It is a shorter route to Europe and western Siberia"

Yes, but SE Asia, including Australia, was settled some time before Europe was. Not so certain about western Siberia though.

"to a lesser extent, to eastern Siberia and E Asia".

Much lesser extent. Remember the Gobi Desert.

"It says that it had to go through a cold region".

Probably, although I don't think you've in any way proved your theory about 'cold' being sufficient.

"It is possible that they arose from K2 after K2 had migrated to SE Asia".

I agree. But no basal K2 is found anywhere other than in and near SE Asia. And its closest relation, K1 or LT, is present in South Asia.

"Why would you think that? [your theory makes it impossible for anyone to have been in SE Asia before the O haplogroups moved south]".

Who do you suggest was in SE Asia before O then?

"EDAR is shown as uncommon in Siberia, where the Mongoloid type is strongest".

Really? Take a look at the wider distribution. It is very common in America and even present in India and northern Europe.

"Too much distance and too much effort required".

There is plenty of time available though.

"It is what most physical anthropologists would have said before the deliverances of modern genetics in the last three decades"

So you're prepared to dismiss modern genetics simply because it doesn't fit what you want to believe?

terryt said...

A further point:

"It is possible that they arose from K2 after K2 had migrated to SE Asia".

And of course it is just as possible that not just K2c and K2d 'arose from K2 after K2 had migrated to SE Asia'. In fact it is extremely likely. Especially when we consider that Wallace's Line lies immediately east of Bali, and in Bali we have K2c. Java lies immediately west of Bali, and in Java we find K2d. And Sumatra lies immediately west of Java, and in Sumatra we find K*. That is a very narrow geographic spread. Further east we find K2b1 immediately across Wallace's line. Is that merely a coincidence? To me it provides considerable evidence for K2b having arisen near Wallace's Line also. K2b is the outsider. Even more so when we consider that K2a is primarily an East Asian clade as well, separated from the other basal K2 groups by the extremely narrow Malay isthmus and the swampy region of Cambodia/Thailand. To me it is quite reasonable to assume that earlier K2 haplogroups regionally intermediate between K2a and the other K2 branches in SE Asia have been replaced by the subsequent expansion of the three O branches.

terryt said...

A little more while I think of it.

A further consideration is that individuals belonging to K2b-P331 but not carrying mutations defining either K2b1-P397 or K2b2-P295 have been recorded in Sumba, Timor, Sulawesi and the Philippines. All of these islands lie east of Wallace's Line, although I'll concede it's arguable as to whether the Philippines lie east or west of the line.

In any case K2b's phylogeny suggests the haplotype reached those islands by sea from Timor/Sumba rather than over land via Borneo. K2b1c-P378 presumably formed from K2b1 in the Philippines as it has been found in the Aeta. Likewise K2b1b-P336 is found in Borneo but is more widespread through eastern Indonesia. M is the most widespread of the K2b1 clades, spread from Malaysia, through eastern Indonesia to Melanesia and New Guinea. It is most probable that the haplogroup's presence at the western end is the product of back movement from at least as far east as Timor, if not further though.

That leaves K2b1a-P405. Individuals lacking the mutations defining any of the several downstream clades have been found in Sumba, in eastern Indonesia. The downstream clades are all centred on New Guinea with P79 also present also in Timor, P60 also present in Australia and S-M230 also present in the Philippines as well as in eastern Indonesia.

To me that all provides strong evidence that K2b2-P295 must have formed somewhere within that geographic spread. K2b1 certainly did, as did K2b.

Gregory76 said...

I asked: "Are you claiming that a route through S Asia and SE Asia goes through more desirable regions than one through C Asia?"
Terry replied: Certainly warmer.

I said: "And ideal conditions are rare".
Terry replied: Especially so through Central Asia, although I do agree it has often provided a route. The route through South Asia certainly would not have included the whole subcontinent. Jungle, desert and the Ganges/Brahmaputra Delta would have restricted the available route.
I reply: That's right: still more disadvantages to your route.

I said: "It is a shorter route to Europe and western Siberia"
Terry replied: Yes, but SE Asia, including Australia, was settled some time before Europe was.
I reply: It is still a shorter route.

I said: "to a lesser extent, to eastern Siberia and E Asia".
Terry said: Much lesser extent. Remember the Gobi Desert.
I reply: There are many ways there that avoid the Gobi.

I said: "It says that it had to go through a cold region".
Terry replied: Probably, although I don't think you've in any way proved your theory about 'cold' being sufficient.
I reply: It's hardly “my theory”; it is that of most physical anthropologists in the recent past. It is an application of Bergmann's rule (or a modified version of it, emphasizing size relative to surface area) and Allens' rule.

I said: "It is possible that they arose from K2 after K2 had migrated to SE Asia".
Terry replied: I agree. But no basal K2 is found anywhere other than in and near SE Asia. And its closest relation, K1 or LT, is present in South Asia.
I reply: And Central Asia.

Terry said: Who do you suggest was in SE Asia before O then?
I reply: C and D.

I said: "EDAR is shown as uncommon in Siberia, where the Mongoloid type is strongest".
Terry replied: ….It is very common in America and even present in India and northern Europe.
I reply: The Mongoloid type is uncommon in the latter 2 areas, and Amerinds are at most diluted Mongoloid, except for Eskimos and Aleuts, and some near-Mongoloids among the Nadene and smaller groups.

I said: "Too much distance and too much effort required".
Terry replied: There is plenty of time available though.
I said: Which does not remove the need for great efforts necessary to cover the great distances.

I said: "It is what most physical anthropologists would have said before the deliverances of modern genetics in the last three decades"
Terry replied: So you're prepared to dismiss modern genetics simply because it doesn't fit what you want to believe?
I reply: My beliefs are the beliefs of most physical anthropologists before the coming of modern genetics. And I have never dismissed modern genetics. But you are dismissing the conclusions of physical anthropologists in the past. We should abandon those conclusions when modern genetics provides compelling reason to do so, but they have not, in this case.

Terry continued: ....That [the distribution of K2c, K2D and K2*] is a very narrow geographic spread....we find K2b1 immediately across Wallace's line. Is that merely a coincidence?
I reply: It is probably 1 of 2 things: (1) a society including K2, K2b, and perhaps others (such as K2b1) migrated to the region, where K2 give rise to K2c, K2d and the K2* samples, and K2b give rise to K2b1; or (2) K2 and K2b1 migrated the region independently: after all, C and D had already migrated there independently of them and probably each other.

Terry said: K2a is primarily an East Asian clade as well,
I reply: As well? K2c and K2d are not East Asian.
Terry continues: ….separated from the other basal K2 groups by the extremely narrow Malay isthmus and the swampy region of Cambodia/Thailand.
I reply: More facts that count against your theory


terryt said...

"It is still a shorter route".

You seem obsessed by this 'shorter route' theory of yours. Easiest route is far more likely to have been a necessity than mere distance. Remember that any early migration would not have a clue about what lay ahead. They would not have set out with a destination in mind. They would have spread through the easiest sections.

"There are many ways there that avoid the Gobi".

And there are many ways through South Asia that avoid the Thar and the Ganges Delta.

"It's hardly 'my theory'; it is that of most physical anthropologists in the recent past. It is an application of Bergmann's rule (or a modified version of it, emphasizing size relative to surface area) and Allens' rule".

I agree that those two rules apply to physical size but that does not explain the majority aspects of the Mongoloid phenotype. The rules only apply to those in the very far north. Most people of Mongoloid phenotype are no different in size to any other population.

"And Central Asia [K2's closest relation, K1 or LT]".

Not so. Only at the very far western end of it, where their presence is easily explained as later exit from S Asia.

"C and D".

Perhaps. But just C1(xC2) and just D2 and possibly the mysterious D* in the Andaman Islands. As well as K2*, K2b, K2c and K2d. I'm not so sure about D though. I suspect its southward movement is related to that of O.

"The Mongoloid type is uncommon in the latter 2 areas"

But definitely present.

"Amerinds are at most diluted Mongoloid, except for Eskimos and Aleuts, and some near-Mongoloids among the Nadene and smaller groups".

I disagree. How many Amerindians have you met? Many are at least as Mongoloid as are SE Asians.

"My beliefs are the beliefs of most physical anthropologists before the coming of modern genetics".

The coming of modern genetics has overturned many early beliefs, even among physical anthropologists. But no for you apparently.

"(1) a society including K2, K2b, and perhaps others (such as K2b1) migrated to the region, where K2 give rise to K2c, K2d and the K2* samples, and K2b give rise to K2b1"

And K2b2. But all the way from Central Asia as an isolated breeding unit? No wonder you have difficulty seeing possibilities for long term movement. It is far more likely that K2 migrated into the region from somewhere very nearby.

"K2 and K2b1 migrated the region independently"

So unlikely as to be unbelievable. K2b2 cannot have formed until after K2b had formed. And by that time K2a, K2c and K2c had also formed. You're postulating that God was in control of some drafting gate, as though humans were sheep after shearing.

"C and D had already migrated there independently of them and probably each other".

Any C2 in SE Asia probably came with O and it is extremely unlikely C1 and D2 entered the region together. D is not present in Australia at all. D appears to have moved into the region considerably after C1 did.

" K2c and K2d are not East Asian".

So you believe SE Asia is separated from East Asia by a minefield and barbed wire?

Gregory76 said...

I said: "It is still a shorter route".
Terry replied: You seem obsessed by this 'shorter route' theory of yours. Easiest route is far more likely...
I reply: The longer the route the greater the chances of dying out.

I said:"There are many ways there that avoid the Gobi".
Terry replied: And there are many ways through South Asia that avoid the Thar and the Ganges Delta.
I reply: Which does not change the fact that the Gobi does not rule out migration to N and E Asia

I said: "It's hardly 'my theory'; it is that of most physical anthropologists in the recent past. It is an application of Bergmann's rule (or a modified version of it, emphasizing size relative to surface area) and Allens' rule".
Terry replied: I agree that those two rules apply to physical size but that does not explain the majority aspects of the Mongoloid phenotype.
I reply: They explain heaviness of build relative to height, proportions of the limbs, extremities and digits to length, facial shape, head shape (to a lesser extent), and flatness of face (including high cheekbones and fatty cheeks). Straight, thick hair is an adaptation to cold.
Terry continues: The rules only apply to those in the very far north. Most people of Mongoloid phenotype are no different in size to any other population.
I reply: the so-called “Classic Mongoloid” type is that of north and central Asia. Even in SE Asia Mongoloids come closer to this than Veddoids and Negritoes.

I said: "And Central Asia [K2's closest relation, K1 or LT]".
Terry replied: Not so. Only at the very far western end of it,
I reply: Which is still in C Asia.
Terry continues: where their presence is easily explained as later exit from S Asia.
I reply: We can easily explain their presence in SW and S Asia by migration from C Asia.

I said: "C and D".
Terry replied: Perhaps. But just C1(xC2) and just D2 and possibly the mysterious D* in the Andaman Islands. As well as K2*, K2b, K2c and K2d. I'm not so sure about D though. I suspect its southward movement is related to that of O.
I reply: How is that relevant to my comment?

I said: "The Mongoloid type is uncommon in the latter 2 areas"
Terry replied: But definitely present.
I said: But still uncommon and so they are not likely to be a place of its origin.

I said: "Amerinds are at most diluted Mongoloid...
Terry replied: I disagree. How many Amerindians have you met?
I reply: I don't base these on my personal acquaintance. Read Imbelloni's article written against the assumption that they are simply Mongoloid.

I reply: "My beliefs are the beliefs of most physical anthropologists before the coming of modern genetics".
Terry replied: The coming of modern genetics has overturned many early beliefs.
I reply: But not all. And it could not overturn the facts about what phenotypes are selected for by various environments.

I said: "(1) a society including K2, K2b, and perhaps others (such as K2b1) migrated to the region, where K2 give rise to K2c, K2d and the K2* samples, and K2b give rise to K2b1"
Terry replied: And K2b2. But all the way from Central Asia as an isolated breeding unit?
I reply: From C or SW Asia, yes. C and D would be there only predecessor.

I said: "C and D had already migrated there independently of them and probably each other".
I replied: Any C2 in SE Asia probably came with O and it is extremely unlikely C1 and D2 entered the region together. D is not present in Australia at all. D appears to have moved into the region considerably after C1 did.
I reply: What is your point?

I said: " K2c and K2d are not East Asian".
Terry replied: So you believe SE Asia is separated from East Asia by a minefield and barbed wire?
I reply: There is no sign of them in E Asia and so they cannot be counted as E Asian to make your point.

terryt said...

"I reply: The longer the route the greater the chances of dying out".

That's a ridiculous comment. The chances of dying out would depend entirely on the suitability of the route, no matter how long it was.

"I reply: Which does not change the fact that the Gobi does not rule out migration to N and E Asia"

I think that round the Gobi is the best explanation for Y-DNA C and D's presence in the east. along with mt-DNA N. But it doesn't fit F's migration east. I'm still unsure about mt-DNA M.

"I reply: How is that relevant to my comment?"

You consistently ignore relationships between haplotypes and their individual distributions. As a result your claims lead to all sorts of inconsistencies.

"I said: But still uncommon and so they are not likely to be a place of its origin".

We're currently having a by-election locally and you sound like a politician. Making claims that are a complete fabrication. If you think I somehow suggest the Mongoloid phenotype 'originated in either of those two regions you need to have your eyes checked.

"Read Imbelloni's article written against the assumption that they are simply Mongoloid".

Again the politician talks. Where did I say they were 'simply Mongoloid'? Surely everyone now accepts they were an admixed population before they even entered America.

"I reply: There is no sign of them in E Asia and so they cannot be counted as E Asian to make your point".

And there's no sign of them in Central Asia and so they cannot be counted as C Asian to make your point.

Gregory76 said...

I said: "The longer the route the greater the chances of dying out".
Terry replied: "The chances of dying out would depend entirely on the suitability of the route, no matter how long it was."
I reply: Certainly not. All else being equal, there is a greater chance of them dying out on a long route than short.

I said: "Which does not change the fact that the Gobi does not rule out migration to N and E Asia"
Terry replied:" I think that round the Gobi is the best explanation for Y-DNA C and D's presence in the east. along with mt-DNA N. But it doesn't fit F's migration east. I'm still unsure about mt-DNA M."
I reply: Why do you think it does not fit the migration of F's descendents but does fit the migration of C, D and N.

Terry said: Who do you suggest was in SE Asia before O then?
I said: "C and D".
Terry replied: Perhaps. But just C1(xC2) and just D2 and possibly the mysterious D* in the Andaman Islands. As well as K2*, K2b, K2c and K2d. I'm not so sure about D though. I suspect its southward movement is related to that of O.
I replied: "How is that relevant to my comment?"
Terry said: "You consistently ignore relationships between haplotypes and their individual distributions. As a result your claims lead to all sorts of inconsistencies."
I reply: Name one.

Terry said: You assume 'cold' is the explanation for the Mongoloid phenotype. The evidence for its region of origin, or at least a substantial portion of the phenotype is here:
http://blogs.discovermagazine.com/gnxp/2013/02/is-girls-generation-the-outcome-of-the-pleistocene-mind/
Check out diagram B.
I said: "EDAR is shown as uncommon in Siberia, where the Mongoloid type is strongest".
Terry replied: ….It is very common in America and even present in India and northern Europe.
I replied: The Mongoloid type is uncommon in the latter 2 areas, and Amerinds are at most diluted Mongoloid, except for Eskimos and Aleuts, and some near-Mongoloids among the Nadene and smaller groups.
Terry replied: But definitely present.
"I said: But still uncommon and so they are not likely to be a place of its origin".
Terry replied: "If you think I somehow suggest the Mongoloid phenotype 'originated in either of those two regions you need to have your eyes checked.
I reply: You brought them up in a discussion of the region of origin of the Mongoloid phenotype.

I said: "Read Imbelloni's article written against the assumption that they are simply Mongoloid".
Terry said: Where did I say they were 'simply Mongoloid'?
I reply: When I said: "Amerinds are at most diluted Mongoloid... “ you said you disagreed, implying that they are undiluted Mongoloid.

Terry said: K2a is primarily an East Asian clade as well,
I reply: As well? K2c and K2d are not East Asian
"I replied: There is no sign of them in E Asia and so they cannot be counted as E Asian to make your point".
Terry said: And there's no sign of them in Central Asia and so they cannot be counted as C Asian to make your point.
I reply: I do not need to suppose them in Central Asia. I merely need to suppose that their ancestors were in SW Asia at some point (which, as a believer in Out of Africa, you must concede).

I say:
Your theory has several problems. It postulates unnecessarily long migration routes, its has difficulty in accounting for the formation of the Mongoloid phenotype, it leads you into misreading the zoogeographic maps and to arguing in a circle, and it posulates a northern migration from Southeast Asia to East Asia when all migrations known from history and archeology have been from East Asia to Southeast Asia. (Yes, you think the migration north occurred long before these migrations South, but the theory that they moved north has all of the above problems.)

terryt said...

You seem to have misunderstood much of what I have said. For example:

"But still uncommon and so they are not likely to be a place of its origin"

I have never claimed America, India or northern Europe to be the place of origin of the EDAR 370A mutation. All the evidence points to northern China or Inner Mongolia to be the region of origin. Its presence elsewhere is the result of movement from that region.

"Amerinds are at most diluted Mongoloid"

All the evidence suggests they are a mix of Mongoloid and Central Asian.

"K2c and K2d are not East Asian"

They are certainly not Siberian. SE Asia is just the most southerly portion of East Asia, and the two regions have probably been genetically connected in many ways since the time of Homo erectus.

"Why do you think it does not fit the migration of F's descendents but does fit the migration of C, D and N".

Because almost all basal F is South Asian or SW Asian, as well some SE Asian. The only exceptions are just two branches: K2a (a little north of the majority of the other Ks), and K2b2. Do you believe in some sort of single expansion from a single small region? The evidence is pointing more and more to the development of Homo sapiens is the result of a very complex series of events, not simply the expansion of a single, suddenly superior population.

"Your theory has several problems. It postulates unnecessarily long migration routes"

I see no problem. Humans have obviously been travelling huge distances as they expanded around the world since H. erectus times.

"its has difficulty in accounting for the formation of the Mongoloid phenotype"

You may not be aware that northern China and Inner Mongolia are both cold and of reasonably high altitude.

"it posulates a northern migration from Southeast Asia to East Asia"

You see a problem with that? The great southern coastal migration, fashionable for a number of years, certainly required that to be so.

"all migrations known from history and archeology have been from East Asia to Southeast Asia".

On what grounds do you assume every series of migrations has been unidirectional? The 'migrations known from history and archeology' are all post the development of agriculture in the Yangtze Valley.

terryt said...

Further to what I see as your reluctance to accept multiple, serial human expansions during our evolution:

Haplogroup phylogeny itself indicates multiple serial expansions, not simple expansions from a single point. We see various haplotype expansions, followed by geographic haplotype diversity, followed by the expansion of a limited sample of haplotypes that have developed within that geographic diversity. K2 provides a perfect example.

In spite of what many such as Maju seem to believe it is extremely unlikely that the five K2 branches had individually formed at the time K2 expanded its geographic range. It is almost certainly so that the individual branches formed within discrete regions some time after K2 had expanded widely. Subsequently just two sub-branches within that diversity were able to spread again independently. Several conclusions are obvious from the K2 phylogeny as we currently understand it.

Firstly, that the string of islands between China and New Guinea were deeply involved in K2's formation.

Secondly, that the expansion occurred at a time when the majority of those islands were connected to the mainland.

Thirdly, that subsequent geographic isolation through rising sea level led to the formation of K2a, K2b, K2c, K2d and K2e.

Fourthly, just as only two sub-branches of the five K2 haplotypes were able to expand into northern Eurasia we find just one sub-branch was able to expand beyond Wallace's Line. K2b1 had certainly developed efficient boats.

Fifthly, the other K2 haplotypes probably hadn't. K2c remained stranded just west of Wallace's Line, in Bali. K2d on the next island to the west, Java. K2* survives on the next island to the west, Sumatra, and on the Malay Peninsula. K2e survives yet further west, in India, presumably having developed at the eastern margin of K2's expansion. Almost certainly originally geographic neighbours to the previous three haplotypes. K2a is today widespread in eastern Eurasia but presumably developed on the northern geographic margin of K2's expansion. Again in close proximity to the other four branches of K2.

Sixthly, we can see that K2a and K2b are the only branches to have spread widely from their region of origin. And K2b has spread far more widely than has K2a. K2b1 spread right through the island east of Wallace's Line, including Australia, although it seems Y-DNA C1b2 had beaten K2b1 across the Line to the Lesser Sunda Islands and to Australia. My bet is that it was with the help of C1b2 that K2b1 was able to improve the boating technology.

What about K2b2? Did they have these more efficient boats? Perhaps it is Wallace's Line that separated K2b1 and K2b2? Such a technology would have allowed K2b2 to move west from Wallacea as rapidly as had K2b1 moved east from the same region. Or even faster, as the region west of Wallacea was already inhabited. The Ganges and Indus Rivers would have provided a rapid entry into SW Asia, from where K2b2 would have been able to enter a largely unpopulated Central Asia. Earlier an extreme northern branch of K2a had reached Ust-Ishim but had become extinct. Expand an diversify yet again.