October 05, 2012

D-statistics reveal contrast between Yoruba and San in "Neandertal ancestry"

I have been exploring the HGDP version released by Patterson et al. (2012) in order to see whether patterns  of "archaic Eurasian" admixture could be detected in living Africans. In a previous experiment, I looked into a surprising link between Denisovans and Africans. Now, I want to investigate possible differences in Neandertal ancestry within Africa itself. An ASHG 2012 abstract suggests that both Neandertal and Denisovan ancestry may be relevant to the African story.

Previous research has concluded that living African groups do not appear to have substantial differences in their apportionment of archaic Eurasian ancestry. This has led to the reasonable idea that the signal of Neandertal admixture in non-Africans was driven by the encounter of Out-of-Africans with a Neandertal population in Asia, perhaps in the Near East, during their early steps outside Africa, involving a single or limited episodes of admixture, although more complex models may be needed as of late.

I have long suspected that part of this signal is due to population structure in Africa itself, and the possibility of archaic admixture in that continent, a hypothesis that is feasible a priori due to the geographical and ecological diversity of Africa and its large surface area, and which has also found support on the basis of recent palaeoanthropological and genetic research. In my opinion, the well-known abundance of polymorphism in Africans vis a vis non-Africans is not only due to the Out-of-Africa bottleneck, but may also be due to an addition of polymorphism via admixture with divergent native African hominin groups.

Advancing a good case for this admixture is rendered difficult by two factors:

  1. The inability of methods relying on linkage disequilibrium to operate on old admixture events, due to the exponential decay of LD over time, which renders archaic-introgressed segments pitifully small at long time scales.
  2. The high temperatures prevalent in sub-Saharan Africa which render DNA preservation problematic, although, to be honest, I have not even seen many attempts to test this hypothesis on whatever prehistoric skeletal remains there do exist from the region.
Why do African groups appear so little different in terms of possible "Neandertal admixture"? I conjecture that the answer lies in the idea that archaic African admixture will tend to even out the signal of Neandertal admixture. To use a geographical analogy, there is little distance difference (in relative terms) between Tokyo and Beijing from the vantage point of New York, but quite a lot from the vantage point or Hong Kong. Tests of archaic admixture rely on relative allele sharing between individuals or populations; consequently, the signal may be muddied by the occurrence of archaic admixture in Africans which -to use our geographical analogy- transposes them from Hong Kong to New York.

Now, consider the Z scores of the D-statistic of the form D(African1, African2, Neander, Outgroup) calculated using different panels and Outgroup being Chimp, Gorilla, or Orang. The raw numbers can be found in this spreadsheet.

Look at the Pearson correlations between the different panels:


While the Z-scores in most of the panels are strongly correlated with each other, the San panel #4 is strongly anti-correlated. An inspection of the raw numbers show why this is the case. For example:


Surprisingly, the San appear more Neandertal-admixed than the Yoruba using all Eurasian and the Yoruba ascertainment, and less so, using the San ascertainment!

A possible explanation for this pattern involves Eurasian back-migration into Africa combined with differential archaic African admixture.

The San may possess Eurasian ancestry consistent with the positive D(San, Yoruba, Neander, Chimp) statistics for all panels except their own; the negative statistics for their own panel is due to their archaic African ancestry which makes them less like Neandertals.

I conjecture that different archaic populations have contributed polymorphism to different African populations.

This question can be addressed empirically on the basis of whole genome sequence data. The Out-of-Africa bottleneck hypothesis suggests that reduced polymorphism in non-Africans is due to loss of variation as a limited number of founders exited Africa, carrying a subset of African variation. If Africans are descended primarily from the modern human groups left behind, then they will all carry the same "missing variation" set not found in Eurasians.

On the other hand, if, as I suggest, modern human groups encountered and admixed with different divergent African hominins, then different African populations will carry substantially disjoint sets of variants, reflecting deep population structure within Africa itself. Time will tell whether this prediction will prove to be true.

31 comments:

Maju said...

How come Green 2010 using similar methods found no significant differences among San and Yoruba in relation with Neanderthal, which is the basis for their finding of Neanderthal admixture in Eurasians, yet you do?

How do you explain that, because in your entry there is not a single sentence that addresses this radical discrepancy.

Isn't your use of related hominids (great apes) distorting the correlation?, because Green and the rest of the Paabo team did not use them at all.

Dienekes said...

How come Green 2010 using similar methods

How about you read and understand a post before commenting on it?

Maju said...

Alright this is my last comment here ever. I had more than enough.

Ted Kandell said...

Diekekes, add the Bonobo too please, if possible?

Also can you try this as I asked with the Hazda as well? I have a feeling there may be more than one African Archaic hominin group involved, separate East, West, and South African Archaic admixture. Even right now we should be able to differentiate different sets of Archaic alleles and admixture within Africa.

I think you are right:
The explanation for the "Denisovan" link to Paleoafricans is a kind of symmetric Archaic relationship to Denisovans via ILS that "skips" the more widespread intervening Neanderthal admixture, but also some possible earlier East Asian (i.e H. erectus) Archaic admixture among Denisovans.

Please also realize one thing:
50% of African PRDM9 zinc-finger alleles which govern recombination hotspot recognition are highly divergent compared to the two 96% modal Eurasian alleles which in fact are functionally equivalent to the Neanderthal-Denisovan alleles. As a result, Africans have totally different and shorter LD than non-Africans or even perhaps Neanderthal-Denisovans.

Would the non-matching recombination hotspots tend to preserve intact longer "archaic" segments among Africans, but appear to greatly reduce LD between African populations?

Dienekes said...

50% of African PRDM9 zinc-finger alleles which govern recombination hotspot recognition are highly divergent compared to the two 96% modal Eurasian alleles which in fact are functionally equivalent to the Neanderthal-Denisovan alleles. As a result, Africans have totally different and shorter LD than non-Africans or even perhaps Neanderthal-Denisovans.


Link?

Diekekes, add the Bonobo too please, if possible?

Also can you try this as I asked with the Hazda as well?


No Bonobo in that set. Also, no Hadza.

Rokus said...

"Why do African groups appear so little different in terms of possible "Neandertal admixture"? I conjecture that the answer lies in the idea that archaic African admixture will tend to even out the signal of Neandertal admixture."

This possibilty should be filtered out somehow, probably using a different method, since rare haplotypes are most unlikely to have a dual origin in Neanderthal AND different African archaic populations.

I already mentioned my own results, that reveal a truly non-ancestral or hybrid, and even 'European' origin of these African Neanderthal admixtures:

Assuming all of Hu's 14% non-Neanderthal non-Denisovan share of Hochreiter's rare haplotypes may be archaic Asian (HSS were just bottlenecked homozygotes after all), and implementing Meyer's 24% lower count for non-African European rare haplotypes compared to non-African Asian rare haplotypes, we have four equations and six variable that can be evaluated for different settings. For instance, you could solve the equations using zeros for each combination of two variables. I arrived at valid results for only three settings. Most peculiar in those results is, that a pronounced shared Afro-European component in all configurations (37%-50%), while the shared Eurasian, Afro-Asian and Afro-Eurasian components are all less pronounced and could even be assumed zero. At a maximum Eurasian share (being the most intuitive configuration) both the ancestral (Afro-Eurasian) and Afro-Asian shared portions approach to zero.

To me the elevated Afro-European portion means a NOT ancestral African link for the Neanderthal rare haplotypes in Africa, but rather the result of Neanderthal-hybrid migrations from Europe to Africa. This must have happened after an Eurasian differentiation event that restructured Neanderthal rare haplotypes into European and Asian forms or configurations. Remarkably, this must have affected ALL Africans.

John Hawks confirmed his team arrived at the same result, albeit using a different method and being unpublished. I figure his team just had to take a closer look at the raw data to find out how the Neanderthal-type rare haplotypes are shared between the continents. We could still do this ourselves for Hochreiter’s 9,554 haplotypes being typified ‘non-exclusive European’, one by one.

Most of all, you should be aware that here we are just talking about the tiny portion of Neanderthal DNA that is different from the current genetic common denominator. This tiny portion is getting very close to the total sum of Eurasian variability. From this point onwards, we could turn around everything and replace the concept of an African HSS reference species by a Neanderthal base + some non-Neanderthal archaic admixtures - actually a complete refutal of the Out of Africa paradigm.

Tobus said...

Surprisingly, the San appear more Neandertal-admixed than the Yoruba using all Eurasian and the Yoruba ascertainment, and less so, using the San ascertainment

the negative statistics for their own panel is due to their archaic African ancestry which makes them less like Neandertals

Again, this is more likely caused by the panels/outgroups you are using. Doing D(San, X, Neandertal, Chimp) with a San panel is guaranteed to make San look less Neandertal than they are.

Some quotes from Patterson et al. 2012 showing the D-statistic needs an ascertainment balanced between the two subject populations:

"The results that follow hold rigorously if we identify the polymorphisms we are studying in an
outgroup
(that is, we select SNPs based on patterns of genetic variation in populations that all
have the same genetic relationship to populations A,B,C). Since only markers with variation in A,B,C are relevant to the analysis, then by ascertaining in an outgroup we ensure that our markers are polymorphic in the root population"

"We assume here that we ascertain in an outgroup (implying polymorphism at the root), and again assume neutrality and that we can ignore recurrent or back mutations"

"We will assume that genetic drift events on distinct edges are uncorrelated, which as mentioned before will be true if we ascertain in an outgroup, and our alleles are neutral"

"Assume that our ascertainment is such that drifts on distinct edges are orthogonal, which will hold true if we ascertained the markers in an outgroup"

"If we ascertain in an outgroup, then if (W,X) and (Y,Z) are clades in the population tree, it is easy to see that E[Numi] = 0."

"More generally, if the relationship of the analyzed populations is as shown in Figure 3c or Figure 3d and we ascertain in an outgroup or in {W,X} then D should be zero up to statistical noise"

"Assume the phylogeny shown in Figure 3b, and that we ascertain in an outgroup to A,B. Then E[D(Chimp, Yoruba; A,B)] = 0"

"D-statistics provide formal tests for whether an unrooted phylogenetic tree applies to the data, assuming that the analyzed SNP are ascertained as polymorphic in a population that is an outgroup to both populations (Y, Z) that make up one of the clades"

Dienekes said...

Again, this is more likely caused by the panels/outgroups you are using. Doing D(San, X, Neandertal, Chimp) with a San panel is guaranteed to make San look less Neandertal than they are.


You are missing the point. It is not about ascertaining in a San or Yoruba panel.

Why would the San appear more Neandertal than Yoruba using a Eurasian (French or Han or Papuan) ascertainment?

Tobus said...

Why would the San appear more Neandertal than Yoruba using a Eurasian (French or Han or Papuan) ascertainment?

Because either a) the San are more Neandertal than the Yoruba (the result is valid), or b) the Yoruba are more Eurasian than the San (the result suffers from bias), or c) there's pre-Neandertal admixture that makes it appear this way (the result is misleading), ...or a combination of any/all three. I will point out that there will be a bias unless the San and Yoruba have the same number of polymorphic alleles in the panel used.

My original point was that the "surprising" negative result for San when using the San panel is to be expected. The result contains a negative bias due to the way the D-statistic works. I'm not saying your theories about archaic admixture are right or wrong, just that you shouldn't take too much from this San result because the inherent bias makes it unreliable.

Dienekes said...

Because either a) the San are more Neandertal than the Yoruba (the result is valid), or b) the Yoruba are more Eurasian than the San (the result suffers from bias), or c) there's pre-Neandertal admixture that makes it appear this way (the result is misleading), ...or a combination of any/all three. I will point out that there will be a bias unless the San and Yoruba have the same number of polymorphic alleles in the panel used.


If the Yoruba are more Eurasian than San (b), then why would the San appear to be more Neandertal than the Yoruba using a Eurasian panel?

As for (c), it is not misleading, because the D-statistic is consistent with either African population structure or Neandertal admixture. It may be misleading only if one always interprets this statistic as evidence of Neandertal admixture (which I don't).

Not sure what you mean about "same number of polymorphic alleles". That is irrelevant when a Eurasian ascertainment is used. And, actually, making the number of polymorphic alleles equal does not remove bias, it introduces it, because different populations may have inherited different amounts of archaic admixture.

Tobus said...

If the Yoruba are more Eurasian than San (b), then why would the San appear to be more Neandertal than the Yoruba using a Eurasian panel?

Because more of the Yoruba alleles would be polymorphic and thus subject to back-mutation. The "ancestral" Neandertal sites in the Yoruba would vastly outnumber any Neandertal sites introduced as part of the European admixture.


Not sure what you mean about "same number of polymorphic alleles". That is irrelevant when a Eurasian ascertainment is used.

What I mean is that there will be a bias if the panel is not neutral - if either population is closer to the panel then it will be subject to the back-mutation bias to whatever the degree of the relationship. The ascertainment used is never irrelevant - it has to be neutral or the results will be skewed.

And, actually, making the number of polymorphic alleles equal does not remove bias, it introduces it, because different populations may have inherited different amounts of archaic admixture.

I don't follow your logic, but the panel must be neutral or you have no chance of getting a reliable answer. If there is other archaic admixture the results will be misleading (but only, as you pointed out, "if one always interprets this statistic as evidence" of admixture).

Dienekes said...

Because more of the Yoruba alleles would be polymorphic and thus subject to back-mutation. The "ancestral" Neandertal sites in the Yoruba would vastly outnumber any Neandertal sites introduced as part of the European admixture.


I am not quite sure I am following your logic.

In fact, there are 15,264 SNPs polymorphic in both the French and San panels, and 14,204 in both the French and Yoruba ones.

So, if you mean that Yoruba and French share more polymorphic sites than San and French, that is not the case.

Tobus said...

In fact, there are 15,264 SNPs polymorphic in both the French and San panels, and 14,204 in both the French and Yoruba ones.

So, if you mean that Yoruba and French share more polymorphic sites than San and French, that is not the case.

I never suggested it was actually the case. It was one hypothetical possibility I was explaining out of 3 hypothetical possibilities I gave to your question of why the San might appear more Neandertal than the Yoruba when using a Eurasian panel. If you are sure the San and Yoruba are equally distant from the panel used, then it's likely one of the other options is the reason - either San are genuinely closer to Neandertals, or the Yoruba have admixture with a pre-Neandertal population making the results misleading, or a combination of both and/or alternative explanations.

My original point still stands - the "surprising" negative results you got for the San when using the San panel are subject to a back-mutation bias and can not be considered accurate.

Dienekes said...

My original point still stands - the "surprising" negative results you got for the San when using the San panel are subject to a back-mutation bias and can not be considered accurate.

I don't know what you think "stands".
When one uses a Han panel on Han, the Han become "more" Neandertal in pairwise comparisons. When one uses a San panel on San, they become "less" Neandertal.

So, whatever bias you seem to think there is does not explain the pattern of the D-statistics.

Some populations become "more Neandertal" when an ascertainment panel from an individual belonging to them is used, and some become "less Neandertal" for the same reason. There is no general "back mutation bias", but rather population-specific behavior that points to some populations having an excess/deficiency of Neandertal admixture, and (in the case of Africans) probably archaic admixture.

German Dziebel said...

@Rokus

"Most of all, you should be aware that here we are just talking about the tiny portion of Neanderthal DNA that is different from the current genetic common denominator. This tiny portion is getting very close to the total sum of Eurasian variability. From this point onwards, we could turn around everything and replace the concept of an African HSS reference species by a Neanderthal base + some non-Neanderthal archaic admixtures - actually a complete refutal of the Out of Africa paradigm."

I think you got it right - that's where it's going. This wouldn't be admxiture anymore, this would be common descent of all living HSS from a "Neandertal-like" population and we still need to identify the HSS homeland, where speciation into HSS from a Neandertal base occurred.

@ Dienekes

"Surprisingly, the San appear more Neandertal-admixed than the Yoruba using all Eurasian and the Yoruba ascertainment..."

Makes sense to me. Compare with the distribution of such a distinctive odontological marker as "shovel-shaped" incisors. They are found at high frequencies in Amerindians and east Asians and then in declining frequencies in West Eurasia, Oceania and Africa. Older dental remains in West Eurasia tend to be richer in shovel-shaped incisors than modern ones, so evolution went from more shoveling to less shoveling. Shoveling in prominent in Neandertals and Asian Homo erectus, too. In Africa, shoveling is significantly more frequent among Khoisan than among Bantu-speakers.

See Morphological and metrical comparison of San and Central Sotho dentitions from southern Africa
by: A. M. Haeussler et al. American Journal of Physical Anthropology, Vol. 78, No. 1. (1989), pp. 115-122.

Tobus said...

When one uses a Han panel on Han, the Han become "more" Neandertal in pairwise comparisons.

I'm sorry, but which Han/Han panel results are you talking about?

So, whatever bias you seem to think there is does not explain the pattern of the D-statistics.

In all the results from your last 3 posts using SNP panels and D-statictics (this post plus here and here) the back-mutation bias explains the pattern of D-statistics perfectly.

There is no general "back mutation bias", but rather population-specific behavior...

The back-mutation bias is a mathematical fact as true as 1 + 1 = 2. Other population-specific factors may also affect the result, and predicting the degree of bias can be difficult when using mixed populations/panels, but in any case where you don't ascertain in an outgroup the D-statistic results are likely to suffer from this bias.

Dienekes said...

I'm sorry, but which Han/Han panel results are you talking about?

Quite simple:

panel1 (French individual ascertainment):

French Han : Neander Chimp 0.0123 2.294

panel2 (Han individual ascertainment):

French Han : Neander Chimp -0.0165 -2.706

So, Han appear more Neandertal using their own ascertainment and less Neandertal using the French ascertainment.

The opposite is true in the case of the San: they appear more Neandertal using OTHERS' ascertainment, and less Neandertal using THEIR OWN ascertainment.

So, it is clear that something population-specific occurs in Africa to cause this behavior.

In all the results from your last 3 posts using SNP panels and D-statictics (this post plus here and here) the back-mutation bias explains the pattern of D-statistics perfectly.

It does not, and certainly not "perfectly", because it does not explain why Africans become less Neandertal-like using their own ascertainments while Eurasians become more Neandertal-like.

Tobus said...

panel1 (French individual ascertainment):
French Han : Neander Chimp 0.0123 2.294

panel2 (Han individual ascertainment):
French Han : Neander Chimp -0.0165 -2.706


The most likely explanation is that there is both Neandertal admixture and back-mutation bias in effect. If there is a bias of 10% and 11% Neandertal admixture in each panel we would expect a positive signal of 1% for each population using it's own ascertainment.


It does not, and certainly not "perfectly", because it does not explain why Africans become less Neandertal-like using their own ascertainments while Eurasians become more Neandertal-like.

The three posts highlighted unusual results for San/Denisova using a San panel, San/Neandertal using a San panel and Africa/Denisova using a Papuan panel. All 3 can be explained by back-mutation bias alone. The only mention of increased Neandertal admixture in Eurasians was when using an African panel, again exactly in line with what we'd expect from the bias.


This is how the bias occurs:
Chimp: AAAAAA
Neand: BBBBBB
Mono: BBBBBB
Poly: BxByBA (50% drift, no admixture)

- The D-statistic will count the final site as ABBA even though it represents random drift not admixture.
- The D-statistic can only give a reliable result if the drift is equal in both populations.
- A panel ascertained in a population will contain more drift for that population than any other.
- Thus a D-statistic using a panel ascertained on one of its subject populations (or a population closer to one of its subject populations) will not give a reliable result.

Rokus said...

This wouldn't be admxiture anymore, this would be common descent of all living HSS from a "Neandertal-like" population and we still need to identify the HSS homeland, where speciation into HSS from a Neandertal base occurred.

I perceive the common denominator with Neanderthal as sufficiently large, and the Neanderthal derived variability as sufficiently wide spread, to doubt there ever was any separate HSS speciation event in any homeland significantly isolated from Neanderthal populations. Some genes or segments must have crossed boundaries by hybridization, though, and acquired predominance by natural selection. Possibly this happened with mtDNA, like in polar bears - but such events don't make a genetic homeland.

eurologist said...

Most of all, you should be aware that here we are just talking about the tiny portion of Neanderthal DNA that is different from the current genetic common denominator. This tiny portion is getting very close to the total sum of Eurasian variability. From this point onwards, we could turn around everything and replace the concept of an African HSS reference species by a Neanderthal base + some non-Neanderthal archaic admixtures - actually a complete refutal of the Out of Africa paradigm.

Rokus,

I don't understand your argument. Diversity has two components: the number of different SNPs, and how widespread they are. Even if Eurasian diversity is hugely impacted by Neanderthal SNPs (which I agree it is) - any single individual only harbors a tiny aspect of that, and a very different one from her next door neighbor, to boot. So, from an individual (not population) perspective, the "base" is still overwhelmingly AMHs.

Dienekes said...

The three posts highlighted unusual results for San/Denisova using a San panel, San/Neandertal using a San panel and Africa/Denisova using a Papuan panel. All 3 can be explained by back-mutation bias alone. The only mention of increased Neandertal admixture in Eurasians was when using an African panel, again exactly in line with what we'd expect from the bias.


I don't know how to respond to this when I just showed evidence that French become more Neandertal than Han using their own panel (French), and Han more Neandertal than French using their own panel (Han).

The exact opposite takes place in Africa, i.e., using the Yoruba panel, the San become more Neandertal-like than the Yoruba, and vice versa.

This is strange behavior indeed which is supposedly "perfectly explained" by your proposed effect.

Tobus said...


I don't know how to respond to this when I just showed evidence that French become more Neandertal than Han using their own panel (French), and Han more Neandertal than French using their own panel (Han).


I guess you didn't understand my explanation. The French don't "become more" Neandertal when using their own panel - they already are more Neandertal when using their own panel. The bias reduces the genuine admixture so it appears the French are only 1% more Neandertal in this particular panel, when in fact they are more so. The same happens when you run Han on the Han panel - more Han-specific Neandertal admixture sites come up, the back-mutation bias waters this down at the non-Neandertal drift sites due and you end up with 1% when in fact, the result should be higher.

Given your propensity for tangents, I should note that this is just one possibility I can see - it's possible (likely) that there are other and better explanations that also work with all the facts. I'm not saying the above is definitely the case, I'm just using it to show that this piece of evidence does not prove there is no back-mutation bias.

The exact opposite takes place in Africa, i.e., using the Yoruba panel, the San become more Neandertal-like than the Yoruba, and vice versa.

You only see it as the "exact opposite" because you are confusing the direction of the bias with the final result, consider:

5 - 10 = -5
15 - 10 = 5

... does this mean in one case I added 10 and in the other I took away 10?

This is strange behavior indeed which is supposedly "perfectly explained" by your proposed effect.

As I've shown repeatedly and you repeatedly ignore, the D-static will exhibit "strange behaviour" unless you ascertain in an outgroup. The bias is not a "proposed" effect, it's a mathematically provable effect that takes about 2 minutes with a pen and paper to understand. I'm happy to explain it again if you genuinely don't get it.

Dienekes said...

I guess we are arguing around in circles. I have proposed an explanation for the empirical pattern observed that involves Neandertal/archaic African admixture. You, on the other hand, propose that the pattern can be "perfectly" described by the "back mutation bias".

I don't really see where you derive your confidence that the pattern is "perfectly" described by the proposed bias. It may very well be, but I'd like to see some argument or data for it that amounts to something more substantial than your 5-10=-5 example.

The bias is not a "proposed" effect, it's a mathematically provable effect that takes about 2 minutes with a pen and paper to understand. I'm happy to explain it again if you genuinely don't get it.

The question is not whether it is a mathematically provable effect or not, but whether it is sufficient to explain the observed pattern.

One can prove with pen and paper whether an effect takes place, but one can only show with data whether this effect explains an empirical pattern.

For example, in a previous post, you contended that while archaic African admixture would have biased the D-statistics in a particular direction, it would require 40x more archaic African than Denisovan admixture to achieve the observed effect. So, while you conceded the mathematical validity of such an effect, you doubted its emprical efficacy and plausibility.

I'm perfectly willing to eat my words, and abandon my conjecture, if someone can show me on the basis of empirical data that the observed deviations are "perfectly" explained on the basis of a particular bias, in which case more complex explanations will be rendered superfluous.

Rokus said...

"Even if Eurasian diversity is hugely impacted by Neanderthal SNPs (which I agree it is) - any single individual only harbors a tiny aspect of that, and a very different one from her next door neighbor, to boot. So, from an individual (not population) perspective, the "base" is still overwhelmingly AMHs."

How do you know that genetic base comes from a non-Neanderthal species, whose anatomy isn't fully ancestral to modern humans anyway? This base is largely shared by Neanderthal, while the differences are tiny and largely preserved in modern variability. Other differences may have been sifted out by natural selection, or replaced by non-Neanderthal introgression.

Maybe it helps to quantify modern variability a bit:

Genotyping [...] 1,131 individuals and 3.1 million single nucleotide variants (SNVs) on chromosome 1 [...] identified 113,963 different rare haplotype clusters marked by tagSNVs that have a minor allele frequency of 5% or less. The rare haplotype clusters comprise 680,904 SNVs; that is 36.1% of the rare variants and 21.5% of all variants. The vast majority of 107,473 haplotype clusters contains Africans, while only 9,554 and 6,933 contain Europeans and Asians, respectively.
[...]
Our findings indicate that the majority of rare haplotypes from chromosome 1 are ancient (Hochreiter et al., 2012)


From Hu we know that such non-African ancient segments are largely Neanderthal:

To gain better understanding in demographic and evolutionary significance of archaic hominin admixture, [...] we identified 410,683 archaic segments in 909 non-African individuals with averaged segment length 83,460bp. In the genealogy of each archaic segment with Neanderthal, Denisovan, African and chimpanzee segments, 77~81% archaic segment coalesced first with Neanderthal, 4~8% coalesced first with Denisovan, and 14% coalesced first with neither (Hu et al., 2012)

I don't see a lot of Eurasian "AMH" variability here. Maybe none, if the remaining 14% could be attributed to archaic Asiatic hominines. What makes you different from your neighbor is most likely derived Neanderthal. Most you have in common could be Neanderthal as well.

Tobus said...

@Dienekes: I guess we are arguing around in circles. I have proposed an explanation for the empirical pattern observed that involves Neandertal/archaic African admixture. You, on the other hand, propose that the pattern can be "perfectly" described by the "back mutation bias".

Yes we are going around in circles and I suspect it's because your idea what I am "proposing" is different to mine. I would have said:

"You have proposed an explanation for the empirical pattern observed that involves Neandertal/archaic African admixture. I, on the other hand, have tried to point out that the empirical pattern observed is flawed due to back-mutation bias"

At the risk of oversimplifying, you are saying "here's a theory on why 1 + 1 = 3", and I'm saying "check your maths".

I hope you can see the difference here. I'm not proposing a different theory in competition to yours, I'm saying that the results are not accurate. If you look at how the D-statistic is calculated, you can see that greater drift in one population over the other will distort the result - that's why Patterson et al. always emphasise that they ascertain in an outgroup. It's not a theory or a proposal, it's the mathematics of the D-statistic. All the observed empirical patterns you are trying to explain are ascertained in one of the subject populations and so will suffer from this distortion to some degree. Trying to explain strange or surprising patterns in these results has no point - the results don't accurately represent the admixture proportion of the subject populations.

I haven't replied to the bulk of your remaining points as all my reponses will be paraphrased repetitions of the above or tangential to the original point I was trying to make. If there is a point in there you still want me to address, please feel free to make it again.


I'm perfectly willing to eat my words, and abandon my conjecture, if someone can show me on the basis of empirical data that the observed deviations are "perfectly" explained on the basis of a particular bias, in which case more complex explanations will be rendered superfluous.

I don't want you to abandon your conjecture - I think it's fits the known data better than any other I've heard and is probably correct. All I'm saying is that when using D-statistics you should be aware that the ascertainment used can artificially distort the result and thus give strange, interesting and surprising but ultimately meaningless results.

Tobus said...

@Rokus: I don't see a lot of Eurasian "AMH" variability here. Maybe none, if the remaining 14% could be attributed to archaic Asiatic hominines

Hu said: we identified 410,683 archaic segments in 909 non-African individuals ... and 14% coalesced first with neither

Are you sure you're not confusing Hu's 14% of archaic segments within the Eurasian genome with 14% of the entire Eurasian "AMH" variablility? From my reading of what you posted they are not the same thing.

Rokus said...

@Tobus:
Unfortunately, I have to piece information together myself, in a preliminary research. Analyzing Hochreiter's rare haplotypes, "the majority" of which his team claimed to be "ancient" and "from times before humans migrated out of Africa", whatever this means, I found the signal for shared Afro-Eurasian or "ancestral" HSS haplotypes to be unconvincing and possibly absent.
Hence, I wondered in a second step that possibly some of the exclusive Eurasian haplotypes could be actually HSS, that nowadays can't be found in Africa anymore. This portion would be at most somewhere in the 14% that doesn't coalesce first with Neanderthal or Denisova. However, I doubt this share will be anything else but Asian archaic.

terryt said...

'This portion would be at most somewhere in the 14% that doesn't coalesce first with Neanderthal or Denisova. However, I doubt this share will be anything else but Asian archaic".

Very likely so, but unfortunately at this stage no-one seems to be looking in that direction for any archaic admixture.

shenandoah said...

I think your stats would be more revealing if you substituted Rhesus Macaque monkey data instead of Orangutan. (I believe Orangutan is actually itself a hybrid species, probably Gorilla x Chimpanzee.) I believe the Rhesus (Macacus) would make a far more relevant comparison.

terryt said...

"I believe Orangutan is actually itself a hybrid species, probably Gorilla x Chimpanzee".

Very unlikely to be so. There is no evidence for chimp or gorilla outside Africa at any time and the orang is an East Asian species. If anything gorillas and chimps are more likely to have descended from something like an orangutan than the other way round. Great apes first arrived in Africa around 15 million years ago, about the time usually taken as the split between the orangutan and chimp/gorilla lines.

shenandoah said...

Terry, England was a Tropical Savannah (like parts of modern Africa) just less than a million years ago. A lot has changed climatically, biologically, and geographically speaking, in 15 million years' time. Some Orangutans possess AB+ blood, which is indicative of hybridization, probably between the Gorilla of which some individuals possess type B+ blood, and the Chimp of which some possess type A blood. The habitable ranges of these animals and of course many others most likely have changed dramatically over the eons.

http://www.independent.co.uk/environment/giant-hippo-bones-reveal-norfolks-ancient-past-as-a-tropical-savannah-6166383.html

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