November 26, 2013

A priori Y chromosome phylogeny from sequencing data

A cool new paper by a team of citizen scientists. The most important new piece of evidence is the joining together of haplogroup M (Papuans) with P in a new MP internal node. Your guess is as good as mine as to whether this MP may have come from, as his descendants are presently spread from the Atlantic via Siberia to the Amazon and all the way to New Guinea. The Mal'ta boy belonged to haplogroup R.

The other interesting discovery is of one Telugu man from India who shares mutations with haplogroups N and O but belongs to neither N nor O, so this defines a new "X" clade in the phylogeny. I am wondering if this could perhaps be called NO0 instead, similar to the way that more basal clades of the entire phylogeny were called A0, A00, and so on? Terminology is tricky...


I am aware of a few commercial ventures to resequence Y chromosomes, and I'm pretty sure that citizen scientists will soon not only be able to re-analyze data such as those from the 1000 Genomes Project, but will be able to generate data of their own.

bioRxiv doi: 10.1101/000802

Generation of high-resolution a priori Y-chromosome phylogenies using “next-generation” sequencing data

Gregory R Magoon et al.

An approach for generating high-resolution a priori maximum parsimony Y-chromosome (“chrY”) phylogenies based on SNP and small INDEL variant data from massively-parallel short-read (“next-generation”) sequencing data is described; the tree-generation methodology produces annotations localizing mutations to individual branches of the tree, along with indications of mutation placement uncertainty in cases for which "no-calls" (through lack of mapped reads or otherwise) at particular site precludes a precise placement of the mutation. The approach leverages careful variant site filtering and a novel iterative reweighting procedure to generate high-accuracy trees while considering variants in regions of chrY that had previously been excluded from analyses based on short-read sequencing data. It is argued that the proposed approach is also superior to previous region-based filtering approaches in that it adapts to the quality of the underlying data and will automatically allow the scope of sites considered to expand as the underlying data quality (e.g. through longer read lengths) improves. Key related issues, including calling of genotypes for the hemizygous chrY, reliability of variant results, read mismappings and "heterozygous" genotype calls, and the mutational stability of different variants are discussed and taken into account. The methodology is demonstrated through application to a dataset consisting of 1292 male samples from diverse populations and haplogroups, with the majority coming from low-coverage sequencing by the 1000 Genomes Project. Application of the tree-generation approach to these data produces a tree involving over 120,000 chrY variant sites (about 45,000 sites if “singletons” are excluded). The utility of this approach in refining the Y-chromosome phylogenetic tree is demonstrated by examining results for several haplogroups. The results indicate a number of new branches on the Y-chromosome phylogenetic tree, many of them subdividing known branches, but also including some that inform the presence of additional levels along the “trunk” of the tree. Finally, opportunities for extensions of this phylogenetic analysis approach to other types of genetic data are examined.

Link

46 comments:

terryt said...

"The most important new piece of evidence is the joining together of haplogroup M (Papuans) with P in a new MP internal node. Your guess is as good as mine as to whether this MP may have come from, as his descendants are presently spread from the Atlantic via Siberia to the Amazon and all the way to New Guinea".

Interestingly: no S. I wonder where it fits in? To me the connection is hardly surprising actually. I have long been reasonably certain that MNOPS originated in Sundaland but its diversification needed to be sorted. And I have recently been toying with the idea that S originated in the extreme SE of Sundaland (Java or Bali) and was first into New Guinea, M in the NE (Borneo or Palawan) and entered Melanesia rather than New Guinea, and NO somewhere near the Burma/China NE India border region. The great mystery has been P. But its apparent connection with M answers a few questions. For example:

"The other interesting discovery is of one Telugu man from India who shares mutations with haplogroups N and O but belongs to neither N nor O"

We already have the interesting case of mt-DNA M connections between Palawan and the east coast of India, exactly where the Telugu live. look at the distribution of M19'53 and M24'41. We also have M42'74 connecting South China, Australia and the east coast of India. Plus the Andaman connections with India. This fits with the 'new' haplogroup's involvement in that movement, probably along with Y-DNA P. Adding to all this is that mt-DNA R's distribution is most easily explained as also having originated in Sundaland and then spread along India's east coast before moving up the main rivers and then northwest out into greater Eurasia.

Hector said...

This should be a comical blow to those who have tried to portray P as strictly and intrinsically "West Eurasian".

Also note that R was present in Siberia ever since Paleolithic times. It was probably not an incursion from the West.

Fanty said...

I recall the first Y-DNA migration maps I ever saw. It technically claimed P gave birth to R almost at the western border of China (or when explained in words it was always the talk about "Central Asia") and that some R would have gone south to India and some would have gone west to Europe (of course with all these old dating like R entering Europe like 20-30K years ago and stuff, beeing the Y-DNA of the Mammoth hunters and all that.

Of course a lot of those old ideas had been trashed but some had been resurrected since. If "Central Asia" would come back as birthplace for "R", Sibiria is closer to its birthplace than Europe is.

Onur said...

From the paper:

"Next, the inclusion of a (relatively rare) haplogroup M sample in this tree (the Papuan, HGDP00542) appears to
have some interesting implications for the phylogenetic structure around haplogroup M. In particular, the tree shows
that the haplogroup M sample shares derived status with haplogroup P samples at several SNP sites, indicating the
existence of an “MP” haplogroup upstream of haplogroup P and haplogroup M, and downstream of haplogroup
K(xLT). (Since no Hg-S samples are considered in the present analysis, the position of Hg-S relative to this new
group cannot be determined here.
)"

Also from the paper:

"Additionally, results for the sample HG03742 (Indian Telugu from the United Kingdom), when compared to other
samples, indicate the existence of a haplogroup (termed Hg-”X” here until a formal nomenclature can be assigned)
upstream of the present haplogroup NO and parallel to the haplogroup “MP” (discussed above) under Hg-K(xLT).
The data for HG03742 indicate that he is ancestral for the haplogroup NO defining SNPs on the current ISOGG Ytree:
M214, P188, P192, P193, P194, and P195. However, he shares a number of mutations with the Hg-NO
samples, as summarized in Table 3, with these mutations defining the Hg-”X” group. As is the case with Hg-”MP”,
the position of Hg-S relative to this new “X” group cannot be determined here.
"

So they did not include any haplogroup S sample in the study. It would be interesting to see in or closer to which clade of K(xLT) S would be positioned if examined: a) the clade of M and P; b) the clade of X* and NO; or c) a separate clade of its own.

eurologist said...

This should be a comical blow to those who have tried to portray P as strictly and intrinsically "West Eurasian".

Hector,

I don't know why anyone would argue that, given its prevalence in India.

However, generally speaking it has a fairly northern and western distribution not only on the subcontinent, but also in Nepal, Iran, and the Caucasus. Many West Asian and European haplogroups (including mtDNA) have their origin in SE Asia - quite obviously there were large back-migrations, likely at times when the the subcontinent had not turned into a desert or semi-desert like during the periods when the monsoon switched off. One of those could have brought MP to the NW of the subcontinent, from which P then moved further North (to the west of the Himalayas) to form Q in central and Eastern Siberia, and R probably just before reaching Siberia (e.g., in Afghanistan). The presence of P there and the large diversity of R in that general vicinity makes it a much more parsimonious birth location than anything east of the Himalayas.

bicicleur said...

I don't understand why we need the new symbol 'X' , we had 'NO' , except none individual NO* was identified yet, and now we have one.

TheXanian said...

Some old theories stated that NO originated from Central Asia. However, I did not buy into those theories. I have always been a fervent advocate of the southern origin of NO.

Now it seems that I'm right. With the identification of X in India, we can finally say that NO most likely had a southern origin, either in India or in Southeast Asia.

Hector said...

There are NO* individuals at very low frequencies in East Asia. An individual with a * y chromosome is not "ancestral", he is as derived as any other. But if *ed y chromosomes are paraphyletic it MAY indicate a place of origin for that haplogroup or haplotype.

The division of East and West Eurasians is an artificial construct. There is nothing intrinsically Eastern or Western and the division is a product of ecology and barriers and the way the continent is shaped.

R for instance is West Eurasian not because it was born as one but because it left more descendent among Westerners than Easterners. It is an accident of demography and population movements than a bona fide pedigree.

And wherever was the place of origin for P haplogroup it seems fairly certain that it had not too distant Southeast Asian ancestors and not too distant Southeast Asian cousins.

Europeans would like to think their ancestors were hunters braving cold weather and gusty winds and already light complexioned and Caucasoid looking.

In reality, this study shows that more likely at that time their (male line) ancestors were sweating under the tropical Sun of Southeast Asia.

Kristiina said...

I change the thread.
Terry, the link with Papuans and yDNA C and mtDNA N is not very strong.
Papua New Guinea Highlands: P1 20/57, P2 10/57, Q1 10/57, Q3 2/57, B 7/57
Papua New Guinea Southern Highlands: K-M9 7%, S 57%, M-M4 36%
Markham: P1 9/67, P2 5/67, Q1 12/67, Q2 10/67, B 22/67
Papua New Guinea Eastern Highlands: C-M38 6%, K-M9 6%, S 47%, M-M4 35%, O1-M119 6%
North Bougainville: P1 4/75, Q1 38/75, B 17/75

In general, the highest Denisovan ancestry map does not match very well with the distribution of yDNA C or mtDNA N(xR). Where and when do you think that yDNA C mixed with the Denisovans? In North China 40-50 kya? Who was in Australia and Sundaland at that time?

If Y-DNA C and mt-DNA N took the northern route already 40-50kya, I would expect them to have arrived first in America, but the only mtDNA N(xR) found is mtDNA X and yDNA C3. As a higher Denisovan ancestry is found in Karitiana, is your idea that C arrived first in America and Q later on?

If Altai and China were inhabited by Denisovans, why yDNA C-rich area in Amur is low on Denisovan ancestry?

It seems that Denisovan ancestry is found in Europeans (for example through Geno2 tests). How did it come to Europe in your opinion?

So is your idea that c. 40-30kya, Papuan MP was in Sundaland, southern Mongoloid NO in Yunnan, Australoid C in North China and Caucasoid D in China and Japan?

Rokus said...

Strange nobody seems to link MP with the migration route that brought Denisovan DNA to Oceania. P developed in northern Eurasia not so far from Mal'ta as well as Denisova Cave, while M is on the other side of the world. Raghavan's supplement Figure SI 9 PCA on MA-1 "using a worldwide SNP panel masked for European and African ancestries" suggests both populations (Mal'ta/European and Oceania) are linked through Central Asia and India rather than East Asia. I wouldn't take the ancestral DNA in southern Asia for granted as the ultimate evidence of origin while this part of the world may rather be a genetic reservoir or refuge. Especially since a South Asian origin of MP would fail to explain Denisovan DNA in Oceania.

Hector said...

Actually I had expected responses like the one from Rokus.

Upon seeing that most readers are resigned to admit the ultimate Southeast Asian ancestry of M526 I was utterly disappointed....

until I saw Rokus' response.

Yeah that is the Eurocentric crowd I remember and am so familiar with.

Does he know that most M526*s are still unresolved and mostly reside in Southeast Asia?

His opinion sounds more like conspiracy theory. At this point there is nothing more pathetic than clinging to an old hope saying "Oh yeah Denisovan connection. Think about that!!!".

Of course nothing is impossible but if anyone wants to dabble in science he has got to know something about the principle of parsimony.

Oh and Kristiina, mtHap A is N(xR) by the way. You see hell a lot of them among East Asians and NA.

terryt said...

"wherever was the place of origin for P haplogroup it seems fairly certain that it had not too distant Southeast Asian ancestors and not too distant Southeast Asian cousins."

Agree 100%.

"In reality, this study shows that more likely at that time their (male line) ancestors were sweating under the tropical Sun of Southeast Asia [Europeans' ancestors]".

That actually becomes even more obvious when we look at mt-DNA R haplogroups. Basically each island in SE Asia and Wallacea has its own representative of the haplogroup. This is not so for either mt-DNA M or N.

"the tree shows that the haplogroup M sample shares derived status with haplogroup P samples at several SNP sites, indicating the existence of an 'MP' haplogroup upstream of haplogroup P and haplogroup M, and downstream of haplogroup K(xLT). (Since no Hg-S samples are considered in the present analysis, the position of Hg-S relative to this new
group cannot be determined here.)"

However I think we can assume that the old KMNOPS-M526 (K(xLT) if you prefer) clade still stands. What this paper does is illuminate the position of P and M within that clade.

"It would be interesting to see in or closer to which clade of K(xLT) S would be positioned if examined: a) the clade of M and P; b) the clade of X* and NO; or c) a separate clade of its own".

I'd be very surprised if it was anything other than c), as is the case for K1-P60, K2-P79 and K3-P261.

"Terry, the link with Papuans and yDNA C and mtDNA N is not very strong".

Perhaps not. But the link between yDNA C and mtDNA N and Australian Aborigines is extremely strong. Australian Aborigines have very little of either Y-DNA MNOPS or mt-DNA M, except in the far northwest where individual clades of Y-DNA C and mt-DNA N are also present.

"Where and when do you think that yDNA C mixed with the Denisovans?"

In the Altai where the genetic structure probably originated.

"Who was in Australia and Sundaland at that time? [40-50 kya]"

No-one in Australia until Y-DNA C and mt-DNA N arrived in Sundaland and rapidly crossed Wallace's Line. But I have become fairly convinced that Sundaland was sparsely populated with Y-DNA F- and mt-DNA M-derived haplogroups before those other lines arrived.

"If Y-DNA C and mt-DNA N took the northern route already 40-50kya, I would expect them to have arrived first in America"

Not necessarily. The climate would probably have prevented them from reaching far enough north. And certainly later extreme cooling would have pushed them further south, allowing the Mal'ta population to later enter an uninhabited northern Eurasia.

"why yDNA C-rich area in Amur is low on Denisovan ancestry?'

Later admixture with one or more non-Denisova populations.

"How did it come to Europe in your opinion?"

With mt-DNA R and Y-DNA P who still had some residue Denisova genetic presence?

"is your idea that c. 40-30kya, Papuan MP was in Sundaland, southern Mongoloid NO in Yunnan, Australoid C in North China and Caucasoid D in China and Japan?"

Not quite. Australoid and North Chinese C were already separate haplogroups. And C must have entered Australia by 45-50 kya, and so must have passed through China long before 40-30kya. At some unknown time D became spread in an arc from the Andaman Islands, through eastern Tibet and across to Japan although it probably didn't actually reach the Andamans until some 10, perhaps 15, kya.

terryt said...

Sorry. A little more for Kristiina:

Don't forget that New Guinea mt-DNA P and B are both N-derived as is Y-DNA C2. According to yopur information P1, P2 and B constitute 37/57 in Papua New Guinea Highlands, 36/67 from Markham and 21/75 of North Bougainville. Now I admit that by the time mt-DNA B entered Papua/New Guinea it would have had minimal Denisova element remaining, but P probably had retained a considerable proportion. Another factor we need to remember is that even in Papua/Australia the level of Denisova presence is hardly 'huge'.

terryt said...

Sorry folks. Me again. Another interesting point in the paper no-one has brought up: look at G. It branches off first from F*. This is no surprise to me at all. In fact I have long assumed it to be the case. F haplogroups 'dropped off' along the way from Africa to South Asia. But what is a surprise here is that South Asian H drops off before IJ. That contradicts the ISOGG phylogeny where IJ is part of IJK and H is a basal haplogroup separately derived from F. Something funny here. And we don't know where to place F1-P91, F2-M427, F3-P96 and F4-M481 in the phylogeny.

eurologist said...

Europeans would like to think their ancestors were hunters braving cold weather and gusty winds and already light complexioned and Caucasoid looking.

Hector,

It's all a matter of time frame. By 45kya (IJ) and ~37-35kya (P, Q, R) European forefathers indeed were braving cold weather and developing Caucasoid features (G also figures into this).

Just for illustration purposes, a fourfold split of MNOPS can be easily envisioned if a group found a nice place to rapidly expand at a river they encountered (e.g., the Brahmaputra-Ganges). Then, one expansion would be south (MP -> M & S ), one north and eventually west below the Himalayas (MP -> P), another one crossed the river and split the same way on the other side: (MP -> NO: NO -> N north predominantly inland, NO -> O south and eventually predominantly coastal E Asia north).

Gregory76 said...

Terry,

I still maintain that KexLT (MNOPS) originated in central Asia, or originated just to the south, in southwest Asia, with most of the descendent lineages originated in central or northern Asia, because so many of their descendents show adaptations to a cold climate. The Mongoloid phenotype seems to have originated with NO, and that phenotype required the very cold climate of the taiga or even the tundra. M and S are found in near Oceania, mainly among Papuans and Melanesians), and seem to me to be the best candidate for the haplotypes of the Murrayians (i.e., those people who contributed the Caucasoid element to Australasia according the trihybrid hypothesis), and Caucasoids from the colder climates could provided the heavyset bodies typically of Melanesians, who in turn contributed to Polynesians and would account for their heavyset bodies, otherwise unusual in Oceania. Finally there is P, ancestor of Q and R. Q seems to have originated in north Asia and acquired a partially Mongoloid phenotype which was retained by those who moved south in the Americas (i.e., the Amerinds), while those who stayed behind becoming fully Mongoloid. R is partially found in the northern regions.

Hector said...

terryt: G drops off first and then H and then IJ then LT

This has been known for quite some time and it is even included in Wiki.

The theory that Denisovans lived only in Altai is weak I think. Their presence in SE Asia is more likely where they were mixed with modern humans.

A lot of work has been done on Y-haplogroup C and C appears to bifurcate into C3 and the rest.
C3 again bifurcates into "Northern" and "Southern".
"Northern" includes NA ones and the majority of Mongols'.

One interesting find is that C1 (found mostly among Japanese) is the closest relative of mysterious European C*s. They together are cousins of Indian C5.
C2 found in SE Asia is the next of Kin and all of them form the "rest" separate from C3.

This work was done primarily by citizen scientists so it will take some time until you see it in a journal.

Kristiina said...

Hector, true, I forgot about A. It is however much more common in the north than in the south. Terry’s model would explain the later strong MA-1 like admixture in America with the arrival of Q, but the time frame is very recent if it happened according to this Mal’ta paper ”at 5 kya, population 4 receives 41% of its genetic material from population 3”. The climate was milder 40-50kya and getting cooler 20-30 kya. If C was well adapted to the northern environment, I do not understand how ”southern” Q would have reached America before C and furthermore just before the last glacial maximum.
Terry, do you think that all mtDNA N, i.e. N(xR) and R and in particular P, came from the north with yDNA C and only mtDNA M came from India? However, Zhong et al. 2010 argued against the norther origin of yDNA C. From the yDNA C tree it does not appear that Indian C is a subset of East Asian C. Do you think that yDNA C split in two, with Indian C taking the southern route and East Asian C taking the northern route?
As it seems that Denisovans were not found in India, someone must have taken the northern route, if we do not claim that the Denisovan-Papuan admixture happened in Indo-China. I still find it somewhat odd that MP would have made it so quickly from Sundaland to Siberia through India, becoming Q, and transformed in c. 5000-10000 years from an Indonesian jungle dweller to a Siberian arctic dweller.

As always, frustratingly, almost all doors remain open.

Daniel Szelkey said...

I looked closely at the tree and the zip file and I noticed some interesting correlations.
First I will discuss the correlations discovered before the paper
-----------------
O1'02 is the ancestor of O1 and O2 but not 03
http://www.phylotree.org/Y/tree/
----------------------------------------
HIJK exists, and has been noted here
http://www.phylotree.org/Y/tree/
------------------------------------
As most of you know according to the paper M'P is the ancestor of M and P.
----------------------------------------
Apparently C1 and C5 belong to the same branch which I call C1'5. C3 does not belong to this branch, I am uncertain about other forms of C.
-----------------------------------
According to http://www.phylotree.org/Y/tree/
F3 is really H2'
but I found something new all F* in India from yfull.com belongs to a separate branch of H which I dub, H3. Note that all FS* from india clustered together, they made one new branch not many branches.
--------------------------
Compared to H3
H1 and H2(formally F3) cluster together and belong to their own group called H1'2
--------------------------------------
NO* does not exist, what was labelled NO* in yfull has been reassigned to a clade of N centred in china, which I labelled N1e.
----------------------------------
R2a1 was not the only clade of R2a. R2a2 is in west India. R2a3 is all over India, one outlier was found which I Labled R2a0 because it looks divergent it was found in Punjab.
---------------------------
E1b1a had a main clade,
but almost all the Mende belonged to a new clade which I labelled E1b1a1m
-----------------------
E1a2 IS PART OF E1a4
------------------------
D2A AND D2B EXIST IN JAPAN D2A IS MUCH LARGER D2B ONLY TAKES UP AROUND 3 PERCENT OF JAPAN D2* WAS NOT PRESENT.
--------------------------------
Q1a1a1 is found in Kinh in Ho Chi Minh City and Peru in Americas, proving not all American Q is the same.
------------------------------------

Naren said...

Telugu man from India who shares mutations with haplogroups N and O but belongs to neither N nor O.

That place has significant Munda(O2) presence and also close to Jwalapuram where they dug for Toba period tools.

There is travel with Java also.

terryt said...

"This [G drops off first] has been known for quite some time and it is even included in Wiki".

Thanks. It doesn't appear in ISOGG though so I wasn't aware of it, although I certainly suspected it to be so.

"The theory that Denisovans lived only in Altai is weak I think. Their presence in SE Asia is more likely where they were mixed with modern humans".

I'm inclined to disagree. SE Asia is a long way from Altai and to me it seems unlikely that the whole population between the two regions would be homozygous for a gene that does not appear to have been taken up much at all.

"This work was done primarily by citizen scientists so it will take some time until you see it in a journal".

Thanks. I look forward to that very much. I am yet to see a study that includes all C haplogroups.

"C appears to bifurcate into C3 and the rest"

That's interesting, and certainly suggests that C's movement east was through Central Asia, not via South Asia.

"Zhong et al. 2010 argued against the norther origin of yDNA C".

Hector's information suggests Zhong is wrong though.

"do you think that all mtDNA N, i.e. N(xR) and R and in particular P, came from the north with yDNA C and only mtDNA M came from India?"

Basically. Although I doubt P, or even R, had formed while N moved. The various N and C subgroups developed locally after the resepective parent haplogroup had populated each separate region. The idea that C/F and M/N used separate routes east certainly explains why both Y- and mt-DNA lines each form two distinct branches. If M and N along with C and F had moved together we would not see such an obvious split in the geographic presence of all four haplogroups.

"As it seems that Denisovans were not found in India, someone must have taken the northern route, if we do not claim that the Denisovan-Papuan admixture happened in Indo-China".

Yes.

"If C was well adapted to the northern environment, I do not understand how 'southern' Q would have reached America before C"

Q had obviously spent some time on the steppe and so had actually become better adapted to the northern environment than had the earlier C. In fact I guess it had been pushed south by cooling climate at some stage.

"I still find it somewhat odd that MP would have made it so quickly from Sundaland to Siberia through India, becoming Q, and transformed in c. 5000-10000 years from an Indonesian jungle dweller to a Siberian arctic dweller".

The movement is by no means particularly rapid. And don't forget it wasn't 'Papuans' who reached the steppe. By the time the haplogroups had reached that region they had become thoroughly admixed with populations they met along the way. That explains the genetic connections between Europeans and South Asians. And probably explains the traces of Papuan ancestry through much of the world.

terryt said...

Sorry. Two posts again:

"It's all a matter of time frame. By 45kya (IJ) and ~37-35kya (P, Q, R) European forefathers indeed were braving cold weather and developing Caucasoid features (G also figures into this)".

Yes. At that point it is not relevant where their deeper ancestry lies. By that time the haplogroups had picked up non-Southeast Asian genes from all over the place.

"Just for illustration purposes, a fourfold split of MNOPS can be easily envisioned if a group found a nice place to rapidly expand at a river they encountered (e.g., the Brahmaputra-Ganges). Then, one expansion would be south (MP -> M & S ), one north and eventually west below the Himalayas (MP -> P), another one crossed the river and split the same way on the other side: (MP -> NO: NO -> N north predominantly inland, NO -> O south and eventually predominantly coastal E Asia north)".

Largely agree. However I think the 'nice place to rapidly expand' was through the islands of SE Asia, when the haplogroup developed (or more likely adopted) the ability to cross Wallace's Line. That let NO and P move back north and reach the Brahmaputra-Ganges, and carry on.

"I still maintain that KexLT (MNOPS) originated in central Asia, or originated just to the south, in southwest Asia, with most of the descendent lineages originated in central or northern Asia, because so many of their descendents show adaptations to a cold climate".

That last comment is perhaps correct for the actual number of descendants, but not so for the number of haplogroups. I agree that most people outside Africa are descended from NO or P but M, S, K1, K2 and K3 are all Sundaland haplogroups, or further east, who show no 'adaptations to a cold climate' at all.

"The Mongoloid phenotype seems to have originated with NO, and that phenotype required the very cold climate of the taiga or even the tundra".

To some extent I agree although C3 and most Q haplogroups also display Mongoloid features so you can't confine the origin to just NO.

"M and S are found in near Oceania, mainly among Papuans and Melanesians), and seem to me to be the best candidate for the haplotypes of the Murrayians (i.e., those people who contributed the Caucasoid element to Australasia according the trihybrid hypothesis)"

Neither of those Y-DNA haplogroups are present in Australia at all, so cannot be the ones who contributed the Caucasoid element to the Murrayians.

"Caucasoids from the colder climates could provided the heavyset bodies typically of Melanesians, who in turn contributed to Polynesians and would account for their heavyset bodies, otherwise unusual in Oceania".

The Polynesian phenotype is usually considered to be the result of cold nights during migrations on the open ocean, not the result of older genes. Melanesians are not actually particularly heavy bodied:

http://www.ncbi.nlm.nih.gov/pubmed/2317002

"Finally there is P, ancestor of Q and R. Q seems to have originated in north Asia and acquired a partially Mongoloid phenotype"

But even within these haplogroups we find many branches which do not have a Mongoloid phenotype. The current paper suggests very strongly that the branches of R and Q that do have a degree of Mongoloid phenotype aquired it through admixture with East Asians who already had it.

Rokus said...

Does he know that most M526*s are still unresolved and mostly reside in Southeast Asia?

Not so long ago it had to be that undersampled Spain was the most diverse source of R1b, but it turned out most they had was a single subclade P312, and the Basques were nothing but a tiny subset of this. Hardly the relict of a unique Amerindian or LGM heritage.
Not so long ago it had to be that undersampled India was the most diverse source of R1a, but all they have - and the rest of Asia as well - is a single subclade Z93. If this wasn't the end of a purported Asian origin of the R1a clade, at least it was the end of the principle of parsimony to achive this goal.
Now enlightenment is expanding, but did't reach far enough to resolve Southeast Asia. And what strong answer the unresolved status of M526 is supposed to supply for Denisovan genes in Oceania?
Apparently the opposite of unpopular evidence is Mysticism? Those desperate for having it their way will soon have to emigrate to another terra incognita.

Onur said...

"It would be interesting to see in or closer to which clade of K(xLT) S would be positioned if examined: a) the clade of M and P; b) the clade of X* and NO; or c) a separate clade of its own".

I'd be very surprised if it was anything other than c), as is the case for K1-P60, K2-P79 and K3-P261.


K1, K2, K3 and K4 too are not included in this study. So we do not know where they would be positioned either if examined.

Gregory76 said...

I said:
"I still maintain that KexLT (MNOPS) originated in central Asia, or originated just to the south, in southwest Asia, with most of the descendent lineages originated in central or northern Asia, because so many of their descendents show adaptations to a cold climate".

Terry replied:
“That last comment is perhaps correct for the actual number of descendants, but not so for the number of haplogroups. I agree that most people outside Africa are descended from NO or P but M, S, K1, K2 and K3 are all Sundaland haplogroups, or further east, who show no 'adaptations to a cold climate' at all.”

I reply:
In fact Melanesians and apparently some Papuans do; as I said later in the post, Melanesians seems to be stocky (and may account for the stockiness of Polynesians).
I would add that Murrayians seem to be the best candidates for the source of stockiness in Melanesians, and that M and S (and I meant to add K) seem to me to be the best candidate for the haplotypes of the Murrayians.

I said:
"Caucasoids from the colder climates could provided the heavyset bodies typically of Melanesians, who in turn contributed to Polynesians and would account for their heavyset bodies, otherwise unusual in Oceania".

Terry replied:
“The Polynesian phenotype is usually considered to be the result of cold nights during migrations on the open ocean, not the result of older genes. Melanesians are not actually particularly heavy bodied:

http://www.ncbi.nlm.nih.gov/pubmed/2317002”

I reply:
I'd be surprised if there was a consensus in favor of that explanation yet. In any case, it seems that if the explanation were true then Micronesians and Melanesians would also be stocky, as they migrated from island to island on the open ocean. Also, I saw no reference in the abstract to the Melanesians.

I said:
"The Mongoloid phenotype seems to have originated with NO, and that phenotype required the very cold climate of the taiga or even the tundra".

Terry replied:
“To some extent I agree although C3 and most Q haplogroups also display Mongoloid features so you can't confine the origin to just NO.”

I reply:
Yes, but that does not affect my point: N and O are still overwhelmingly Mongoloid, which requires an origin in very cold and therefore far northern climates.


I said:
"Finally there is P, ancestor of Q and R. Q seems to have originated in north Asia and acquired a partially Mongoloid phenotype"

Terry replied:
“But even within these haplogroups we find many branches which do not have a Mongoloid phenotype. The current paper suggests very strongly that the branches of R and Q that do have a degree of Mongoloid phenotype aquired it through admixture with East Asians who already had it.”

I reply:
The Mongoloid phenotype could not originate in East Asia, because it is not cold enough, as is shown by the fact that in similar climates to the west, such as those in Europe, Mongoloids are not native. Even Mongolia is not cold enough, because in lands just to the west Caucasoids were there first.
That leaves only North Asia as cold enough for the Mongoloid cradle.
However, North China is probably cold enough for the Murrayians, who are not quite so cold adapted, being taller than classical Mongoloids.

terryt said...

"the zip file and I noticed some interesting correlations".

Brilliant. Although I noticed several omissions. Such as:

"K1, K2, K3 and K4 too are not included in this study. So we do not know where they would be positioned either if examined".

Some of those no longer exist but Melanesian K-P79 appears as K3 in the Phylotree (K2 in ISOGG). But I see no mention of ISOGG's K1-P60 or K3-P261 in the Phylotree version. Can you place them Daniel?

"F3 is really H2' but I found something new all F* in India from yfull.com belongs to a separate branch of H which I dub, H3. Note that all FS* from india clustered together, they made one new branch not many branches.

Really interesting, and makes sense.

"That place has significant Munda(O2) presence"

The Munda language is 'Austro-Asiatic' and both it and Y-DNA O2 almost certainly entered India from the east. As did mt-DNA M42, closely related to an Australian Aborigine haplogroup.

"Now enlightenment is expanding, but did't reach far enough to resolve Southeast Asia".

Southeast Asia is quite well resolved as it was assumed to provide connections between East Asia and America, where all this started. That's why mt-DNAs start with American haplogroups A, B, C and D. Next we have SE Asian E and F, East Asian G and, finally, South Asian H. Y-DNA may be a little less resolved although Daniel's comment shows improvement is possible. His link certainly places K3, M, NO, P and S in a single clade. And we still have SE Asian/Papuan/Australian K1-P60 and K3-P261 to add in somewhere.

terryt said...

"I'd be surprised if there was a consensus in favor of that explanation yet [The Polynesian phenotype is usually considered to be the result of cold nights during migrations on the open ocean]".

Be surprised. The explanation has been totally accepted for something like 40 years. What do you mean by 'yet'? The paper is from 1990. Here is a more extended version by the author:

http://www.jps.auckland.ac.nz/document/?wid=3029

"Melanesians seems to be stocky (and may account for the stockiness of Polynesians)".

And I tried to point out, 'Melanesians are not actually particularly heavy bodied'. This quote from the above link:

"Rhoads (1984) offers an interesting statistical analysis of anthropometric data from several Solomon Islands groups, which make apparent that two populations, the Lau of northern Malaita and the Ontong Javanese, are particularly heavy. For these same two groups the functionally related nasal height and chest breadth also are particularly large. It is just these two groups whose environment and economy are, or were, essentially maritime, in contrast withthe land-based economies of all the other groups, save one".

So these two Melanesian groups stand out from the others precisely because they are 'particularly heavy'.

"if the explanation were true then Micronesians and Melanesians would also be stocky, as they migrated from island to island on the open ocean".

No. Melanesians are mostly land-based. You will se reference to Micronesians in the above link.

"I would add that Murrayians seem to be the best candidates for the source of stockiness in Melanesians, and that M and S (and I meant to add K) seem to me to be the best candidate for the haplotypes of the Murrayians".

You're ignoring what I pointed out before, 'Neither of those Y-DNA haplogroups are present in Australia at all, so cannot be the ones who contributed the Caucasoid element to the Murrayians'.

"In fact Melanesians and apparently some Papuans do"

Yes. M and S are confined to New Guinea and the islands of Melanesia. So how can either M or S be responsible for any 'haplotypes of the Murrayians'. And the only K haplogroup Australia and New Guinea have in common is K1-P60. Within Australia that haplogroup is primarily confined to the northwest, nowhere near the Murray Basin. In case you don't realise it the Murray River is in southeastern Australia:

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

My brother and his son farm in the Murray basin just south of Echuca. That basin is where the 'Murrayans' are from. They are not present at all in Melanesia. The one haplogroup difference they have compared to other Australian Aborigines is an elevated level of mt-DNA M42.

"N and O are still overwhelmingly Mongoloid, which requires an origin in very cold and therefore far northern climates".

I point out at the other post that 'cold alone' is not sufficient to explain the Mongoloid phenotype.

"The Mongoloid phenotype could not originate in East Asia, because it is not cold enough, as is shown by the fact that in similar climates to the west, such as those in Europe, Mongoloids are not native".

But Europe is much more low-lying than is the Siberian Plateau.

"Even Mongolia is not cold enough, because in lands just to the west Caucasoids were there first".

Yes. And they have since been replaced there by a better-adapted Mongoloid phenotype.

"North China is probably cold enough for the Murrayians, who are not quite so cold adapted, being taller than classical Mongoloids".

What do you actually mean by the term 'Murrayians'?

terryt said...

A few more observations regarding Daniel's list:

"H1 and H2(formally F3) cluster together and belong to their own group called H1'2"

That explains my earlier confusion, 'But what is a surprise here is that South Asian H drops off before IJ'. Just like G and IJK, H formed in SW Asia and just the derived H1 entered South Asia. As did the derived K(xIJ).

"all F* in India from yfull.com belongs to a separate branch of H which I dub, H3. Note that all FS* from india clustered together, they made one new branch not many branches".

The single F (H3) branch also entered South Asia along with the derived H1 and K(xIJ). So just one of the four basal 'F' branches can be considered 'South Asia'. The remainder are spread through the Anatolian/Iranian Plateau. No hint of a 'coastal migration' even.

"Apparently C1 and C5 belong to the same branch which I call C1'5. C3 does not belong to this branch, I am uncertain about other forms of C".

I have just found a piece of paper with some information on C's phylogeny, probably from Ebizur. In this tree C3 does belong with C1 and C5 though. C forms three groups: The C1/C3/C5 in Central/East Eurasia, a C2/C* in SE Asia/Wallacea and the remaining Australian C4 on its own. Makes sense except for the apparent sequence of divergence. Unless C miraculously leapt from somewhere near Africa all the way to Australia without leaving a trace of its passing.

Ebizur said...

"H1 and H2(formally F3) cluster together and belong to their own group called H1'2"

This is slightly inaccurate. The authors of the present study have stated explicitly, "The present study lacks F3-M282 samples." They also note, "A recent study reported that F3 is also a branch of M3035," citing van Oven, M., et al., "Seeing the Wood for the Trees: A Minimal Reference Phylogeny for the Human Y Chromosome." Human Mutation, 2013 (online tree http://www.phylotree.org/Y).

It is clear now, however, that (former) H-M69, (former) F3-M282, and (former) South Asian F*-M89(xG-M201, H-M69, I-M170, J-M304, K-M9) all belong to a monophyletic clade marked by at least 28 SNPs (M3035, M3076/Z4354, M2942/Z4276, M2896/Z4251, M2713/Z4171, Z4309, Z4166, M3095, M3070/Z4346, M3062/Z4340, M3058/Z4338, M3052/Z4336, M3035/Z4329, M3010/Z4318, M3004, M2992/Z4307, M2966/Z4292, M2958, M2957, M2955/Z4282, M2945/Z4278, M2939/L901, M2938, M2936/Z4273, M2920/Z4265, M2856, M2826/Z4221, M2773). This clade may be called haplogroup H-M3035 or H-L901, with the former H-M69 demoted to H1-M69. Assuming an average of 90 years for the accumulation of each of the 28 SNPs known to distinguish the H-M3035 clade from IJK-M522/M523, we must consider that at least 2,520 years of "proto-H" evolution have passed after the split of proto-H from HIJK-M578 and before the MRCA of all extant representatives of H-M3035.

It also has been shown that all (or at least the majority) of the former H-M69(xH1-M52, H2-Apt) Y-DNA from South Asia forms a clade that contains H2-Apt as opposed to H1-M52. In other words, the former H*-M69 representatives from South Asia have turned out not to be a polyphyletic assortment of various subclades of H-M69, but rather a sort of "proto-H2" or early branches off the stem that has led also to (former) H2-Apt after the split from (proto-)H1-M52.

Returning to the topic of F3-M282, it is important to note that it is not yet clear how this clade relates to the division between H-M69 (new H1-M69) and H*-M3035(xH1-M69). It is clear, of course, that F3-M282 is upstream of H1-M69, but it is not clear whether F3-M282 should turn out to be closer to H1-M69, closer to H*-M3035(xH1-M69), or closer to neither (i.e. one branch of a true trifurcation below H-M3035).

Ebizur said...

I apologize for the double post, but this is a continuation of my previous comment:

Looking beyond the newly discovered H-M3035, it is known that F2-M427/M428, which has been found in some populations (especially Lolo-Burmese) of southwestern China, is upstream of HIJK-M578. The authors of the present study have noted also that "the tree shows that results for the 1000 Genomes Project sample HG02040 (Kinh population) indicate a separation in SNPs defining the branch for haplogroup F (Hg-F). In particular, although HG02040 has most of the key SNPs in common with other Hg-F samples, his results indicate ancestral results for three SNPs that are found to be derived (or no-call) for all other Hg-F samples (including Hg-G, Hg-H, Hg-I, and Hg-J): 8589031 C>T (variously known as F1329, M3658, PF2622, and YSC0001299), 14367181 G>A (known as M3684 and PF2661), and 22475403 A>C (which we denote here as Z12203); there is also another SNP with similar pattern, 14237670 C>T (known as M3680 and PF2657), where the tree indicates a back-mutation in the Hg-O sample NA18994. Thus, it would appear that these three SNP mutations define a new haplogroup (which might be called haplogroup “GHIJK”) that is downstream of Hg-F and upstream of the recently-identified Hg-HIJK [3]. Due to the presence of no-calls at key sites in HG02040, further study of this sample could identify additional “approximate Hg-F” SNPs for which HG02040 is ancestral (or alternatively, there is also a remote possibility that these ancestral results in HG02040 could be shown to be spurious)." In other words, there should be another basal branch of F-M89 in Vietnam (the Kinh are the mainstream/general population of Vietnam, i.e. the population of native speakers of the Vietnamese language) unless this F*-M89 individual's unusual results are spurious or he actually belongs to F2-M427/M428.

terryt wrote,

"I have just found a piece of paper with some information on C's phylogeny, probably from Ebizur. In this tree C3 does belong with C1 and C5 though. C forms three groups: The C1/C3/C5 in Central/East Eurasia, a C2/C* in SE Asia/Wallacea and the remaining Australian C4 on its own. Makes sense except for the apparent sequence of divergence. Unless C miraculously leapt from somewhere near Africa all the way to Australia without leaving a trace of its passing."

I do not recall having written about any such finding on the basis of SNPs. Early work on Y-STRs of members of C-M130 appeared to show a clear split between C2-M38 and C*-M130(xC1-M8, C2-M38, C3-M217, C4-M347, C5-M356) on the one hand and C1-M8, C5-M356, and C3-M217 on the other, with C4-M347 having an ambiguous intermediate or central position. However, it recently has been discovered that an individual from New Zealand who is M130+, M38+, and M217- shares the K29 SNP with members of C-CTS11043, which includes C1-M8 and the recently discovered European C6-V20, and C5-M356. The present study has confirmed that C1-M8 and C5-M356 form a monophyletic clade relative to C3-M217, but it has not included any representatives of C2-M38, C4-M347, or C6-V20.

Considering the SNP and STR data together, I would guess that C-M130 has expanded rapidly and branched off in the following order: {C3-M217 {C2-M38 {C5-M356 {C1-M8 C6-V20}}}}. I am unsure about C4-M347 and extant C*-M130.

A deep division within C3-M217 also has been apparent since the earliest studies of Y-STRs, with one branch leading to the types of C3-M217 common in East Asians (and extending into Southeast Asia and the Himalayas, e.g. Vietnamese and Burushos) and another branch leading to the types of C3-M217 common in North Asia (including some peoples at the margin, like Mongols) and America. Several SNPs that support this structure have recently been found, so I expect that the story of the evolution of C3-M217 should gain a great deal of detail soon.

Ebizur said...

I forgot to explicitly mention the remaining possibility for the relationship among F3-M282, H1-M69, and H*-M3035(xH1-M69): namely, that H1-M69 and H*-M3035(xM69, M282) will be found to coalesce with each other prior to coalescing with F3-M282. In other words, the possibility that H-M69 and H-M3035(xM69, M282) will turn into H1a-M69 and H1b-???, and F3-M282 will turn into H2-M282. This result would be the most natural when one considers the geographic distributions of the clades in question, with H-M282 being found in West Asia and Europe, and H-M69 and H-M3035(xM69, M282) being found in South Asia.

Hector said...

Terryt says
"... the branches of R and Q that do have a degree of Mongoloid phenotype aquired it through admixture with East Asians who already had it."

But R and Q's Caucasoid phenotypes (if they have it) are most certainly acquired through admixture as well.

C2 and C1 C5 C6 share F3393/K29 which C3 lacks so your tree cannot be right. K system was developed by Mr Jeong who is, as far as I know, a Korean graduate student in a different field. He funded the F3393 test on a C2 man who had no interest whatsoever.

It seems that ebizur knows much about him; they probably know each other. I wonder whether Jeong knows who is ebizur and his comical anti-Korean hysteria.

Daniel Szelkey said...

The authors noted a lack of F3 because they did not realize that F3 was really H2. If you look carefully at the phlyogenic tree at http://www.phylotree.org/Y/tree/
and the revision history of haplogroup F on wikipedia, you will not that both are the same clade defined by P96. @Terryt I have no data for K1, K2,K3,K4 because they were not studied if you want it is easy to downlaod the supplementary zip file on most windows computers

Daniel Szelkey said...

Also, any paper on C'S phylogeny from Ebizur would be highly appreciated.

Rokus said...

an individual from New Zealand who is M130+, M38+, and M217- shares the K29 SNP
How come that K29 was not found in C2 before? If true, this would unite C1 (Japan), C5 (~India) and C6 (Europe) together with C2 (~Oceania).
Australian C4 is still suspected to make a Holocene cluster with Indian C* (Redd 2002, Pugach 2013), what is a feat since this would imply considerable Y-DNA replacement from a single Indian source not so very long ago, against 11% contemporaneous Indian admixture. I wonder if Indian C* and Australian C4 are truly different from the disparate C-K29 branch or just in need of more investigation.
Even without excluding C4 as a reliable ancient marker in Oceania, C3 against CXC3 might be indicative of an originally northern ancestral branching that may be equally reminiscent in MP?

Gregory76 said...

I said:
"N and O are still overwhelmingly Mongoloid, which requires an origin in very cold and therefore far northern climates".

Terry said:
“I point out at the other post that 'cold alone' is not sufficient to explain the Mongoloid phenotype.”

I reply:
The question at this moment is not whether cold is sufficient to explain it but whether cold is necessary to explain it, and it is necessary.

I said:
"The Mongoloid phenotype could not originate in East Asia, because it is not cold enough, as is shown by the fact that in similar climates to the west, such as those in Europe, Mongoloids are not native".

Terry replied:
“But Europe is much more low-lying than is the Siberian Plateau.”

I reply:
Yes, Siberia—that's where I say that the Mongoloid phenotype originated.

I said:
"Even Mongolia is not cold enough, because in lands just to the west Caucasoids were there first".

Terry said:
“Yes. And they have since been replaced there by a better-adapted Mongoloid phenotype.”

I reply:
Yes, but they were replaced by invasion or at least migration by Mongoloids coming in—not by evolution from a Caucasoid phenotype under the influence of natural selection.

I said:
"North China is probably cold enough for the Murrayians, who are not quite so cold adapted, being taller than classical Mongoloids".

Terry asked:
“What do you actually mean by the term 'Murrayians'?”

I reply:
I am using “Murrayian” in the way J. B. Birdsell did in his trihybrid hypothesis as to the ancestry of Australians and Papuans:
Murrayians are the Caucasoid element,
Carpentarians are the Veddoid element
Tasmanoids are the Negritoid element
Carpentarians are stronger in Australia and Tasmanoids in New Guinea, with Murrayians a strong minority in each region.
I thought the hypothesis was unnecessarily complex, until the genetic discoveries of recent decades. It turns out that male C and female N are stronger in Australia, male M and female Q stronger in New Guinea, and male K and female P are a strong minority in each region .
So it seems that
male C and female N could be Carpentarian
male K and female P could be the Murrayian
and
male M and female Q could be Negritoid,
except that since K is a direct descendent of M, I suspect that they are Murrayians, too, and they killed the original husbands of the Qs and married the widows.
(Now some Ms have been separated from M as S, but I expect, as regards this hypothesis, what is true of M is true of S.)
Also, I carry no brief for the names: “Murrayian” and “Carpentarian”, or their geographical implications; we can call them "Caucasoid" and "Veddoid".

terryt said...

"I do not recall having written about any such finding on the basis of SNPs".

Sorry. I probably just assumed it was you as I take special note of any comments you make on the subject of Y-DNA haplogroup connections.

"I would guess that C-M130 has expanded rapidly and branched off in the following order: {C3-M217 {C2-M38 {C5-M356 {C1-M8 C6-V20}}}}. I am unsure about C4-M347 and extant C*-M130. A deep division within C3-M217 also has been apparent since the earliest studies of Y-STRs, with one branch leading to the types of C3-M217 common in East Asians (and extending into Southeast Asia and the Himalayas, e.g. Vietnamese and Burushos) and another branch leading to the types of C3-M217 common in North Asia"

That places C's origin firmly in a reasonably northern region of East Asia, surely. And even F may have an eastern origin:

"It is clear, of course, that F3-M282 is upstream of H1-M69, but it is not clear whether F3-M282 should turn out to be closer to H1-M69, closer to H*-M3035(xH1-M69), or closer to neither (i.e. one branch of a true trifurcation below H-M3035)".

Thanks for that clarification.

"or alternatively, there is also a remote possibility that these ancestral results in HG02040 could be shown to be spurious"

We await further research with interest!

terryt said...

"But R and Q's Caucasoid phenotypes (if they have it) are most certainly acquired through admixture as well".

True. We all know that haplotypes are just one of many genetic markers.

"Australian C4 is still suspected to make a Holocene cluster with Indian C* (Redd 2002, Pugach 2013)"

Actually C* is far more likely to be Southeast Asian in origin rather than Indian. It is a hangover from the 'great southern migration theory' that places it in 'India'.

"Even without excluding C4 as a reliable ancient marker in Oceania"

C4 is not really 'Oceanian'. It is completely confined to Australia. C2, especially C2a, is the Oceanian branch of C.

"I am using 'Murrayian' in the way J. B. Birdsell did in his trihybrid hypothesis as to the ancestry of Australians and Papuans"

You seem confused by his usage though. For Birdsell 'Murrayian' defined the type present in the Murray Basin. You appear to be totally ignoring those people in your consideration. For example:

"male K and female P could be the Murrayian"

Neither are common in the Murray Basin. Y-DNA C4 and mt-DNA M42 are the most prominent, although specifically P4b is present in the same proportion as is M42.

"male C and female N could be Carpentarian"

Except that these haplogroups are in no way characteristic of thwe Carpentaria region. And don't forget that P is part of R which is part of N.

"male M and female Q could be Negritoid"

Except that these two haplogroups are not present at all in Australia as far as we know and Negritos are absent from the New Guinea/Australia region.

"except that since K is a direct descendent of M, I suspect that they are Murrayians, too, and they killed the original husbands of the Qs and married the widows".

Q is part of M29'Q and most probably entered New Guinea/Melanesia with K-derived haplogroups, probably S, followed by M. C in New Guinea is entirely C2, confined to the 'Bird's Head' and northern coastal regions. Probaly associated in some way with the Austronesian expansion.

"Also, I carry no brief for the names: 'Murrayian' and 'Carpentarian', or their geographical implications; we can call them 'Caucasoid' and 'Veddoid'".

The clearest connection with 'Veddoid' would be mt-DNA M42. Apart from that it is very difficult to see any connection between Australia and South Asia.

terryt said...

Sorry. One more thing:

"The question at this moment is not whether cold is sufficient to explain it but whether cold is necessary to explain it, and it is necessary".

Yes. And Inner Mongolia is certainly very cold in Winter:

http://en.wikipedia.org/wiki/Inner_Mongolia#Climate

Quote:

"The winters in Inner Mongolia are very long, cold, and dry with frequent blizzards, though snowfall is so light that Inner Mongolia has no modern glaciers[16] even on the highest Helan peaks".

Rokus said...

Actually C* is far more likely to be Southeast Asian in origin rather than Indian. It is a hangover from the 'great southern migration theory' that places it in 'India'
I rather think Indian C* is a hangover from previous assessments that still assigned 2% to this grouping without being conscient of C5. I wonder what Southeast Asia has to do with it. Most C has already been resolved as either C3 or C-K29. I understand the K29 test is quite new and badly in need of being applied to resolve the last bit of Indian C*.

terryt said...

"I looked closely at the tree and the zip file and I noticed some interesting correlations".

I'm not sure if I missed it the first time I looked at the tree or if its been modified. C does show the phylogeny several have been discussing here. But to me the most interesting 'new(?)' bit is the pairing of NO and S. I certainly didn't notice that before. So, instead of the situation some were hoping for that M and S would prove to belong to a single clade the phylogeny now shows them to be as different as it's possble to be within MNOPS.

"I rather think Indian C* is a hangover from previous assessments that still assigned 2% to this grouping without being conscient of C5. I wonder what Southeast Asia has to do with it. Most C has already been resolved as either C3 or C-K29. I understand the K29 test is quite new and badly in need of being applied to resolve the last bit of Indian C*".

I think C in India has long been tested for C5 or C3 but there has always been a proportion that is neither. Ebizur posted this commnet earlier:

"Early work on Y-STRs of members of C-M130 appeared to show a clear split between C2-M38 and C*-M130(xC1-M8, C2-M38, C3-M217, C4-M347, C5-M356) on the one hand and C1-M8, C5-M356, and C3-M217 on the other, with C4-M347 having an ambiguous intermediate or central position".

So C* (in the form of C*-M130(xC1-M8, C2-M38, C3-M217, C4-M347, C5-M356)) presumably stiil exists. As well as South Asia Wikipedia has always placed C* in SE Asia:

http://en.wikipedia.org/wiki/Haplogroup_C-M130

Of course Wiki is not the most reliable of sources. In fact it still suggests that C 'represents a great coastal migration along Southern Asia, into Southeast Asia and Australia, and up the East Asian coast'. But I quote:

"Patrilines that belong to Haplogroup C-M130 but do not belong to any of its identified subgroups are labeled as Haplogroup C*, which are found at low frequencies along the southern coast of Asia from India to Vietnam and into the interior of Yunnan province in southwestern China, Indonesia, and Micronesia".

That means C* is basically found right around the South China Sea. We know from various pieces of evidence that people have moved from that region both east into the Pacific and west into India and so C*'s very minor presence in India is most probably the result of such movement from SE Asia, not a remnant of any early C movement through South Asia.

Hector said...

I did not see pairing of NO and S in this paper. That pairing only happens in Chris Morley's tree and he made some wild guesses in my opinion. The sample has many problems like showing L781+ M45-.

terryt said...

@ Hector:

True.

Chris Morley said...

Hector,

My experimental phylogenies are computer-generated. I don't feel it is appropriate to say I made "wild guesses".

On the contrary, I have from the beginning stressed that my algorithm's identification of an "NOS" node is tentative. See the K(xLT)-1 level of any of my experimental phylogenies since the first version (for example, line 287 of the 8 November 2013 version). The footnote for this line is prefaced by "[i]f the classification of kit 213359 is correct". Followed by "further testing is needed to confirm kit 213359’s membership in S-M226".

Regarding L781, 2219 of 2223 Geno 2.0 kits in my collection were reported L781+ on the FTDNA Y-SNP report pages. So you shouldn't draw any conclusions from the fact that the possible haplogroup S kit was scored L781+. On line 1 of my experimental phylogenies I suggest that L781 may be consistently mis-called.

Lathdrinor said...

How did the argument over P *ever* become West Eurasia vs. Southeast Asia? P's virtual absence in East Asia and Eastern Europe argues against an East/Southeast Asian emergence, on one hand, and on an Eastern European steppe emergence, on the other. Western Europe is of course even less plausible.

Standard theories about P-M45 indicate that it emerged in Central Asia. MP creates problems for this narrative mainly because M is found primarily in the southeast corner of the world, so to speak. But the presence of P in South Asia ought to serve as a bridge.

The expanse between Central and South Asia looks parsimonious for the homeland of P. I fail, however, to understand the bringing up of racial types for P without attendant Paleolithic data. It's too bad Tianyuan failed to produce a Y result.

terryt said...

Thanks for the clarification Chris.

"Standard theories about P-M45 indicate that it emerged in Central Asia. MP creates problems for this narrative mainly because M is found primarily in the southeast corner of the world, so to speak".

It is not just the discovery of MP that creates problems. The whole MNOPS group is the problem. That grouping makes 2 of the 4 haplogroups Sunda, and NO could easily be Sunda as well. That's 3 of 4. Not to forget about the three K haplogroups that ISOGG currently has as being part of K(xLT), either a single grouping within K(xLT) or three more separate groups within MNOPS. We have K1-P60 in Australia and New Guinea, K2-P79 in Melanesia and K3-P261 in Bali (Chris has just K-P79 in his tree, called 'K3' there, but has its position relative to MP and NO doubtful). All Sunda or further east. That leaves just P outside Sunda.

"But the presence of P in South Asia ought to serve as a bridge".

Indicating its route west from Sunda?

"I fail, however, to understand the bringing up of racial types for P without attendant Paleolithic data".

We all know (or should know)haplogroups are a poor indicator of 'race', as is shown by the recent Denisova type found in Spain who looks 'proto-Neanderthaloid'.

"It's too bad Tianyuan failed to produce a Y result".

We know the mt-DNA was B though, whose origins are fairly obviously further south, possibly near Sunda. It would not be at all surprising if NO had made it that far north by the time of Tianyuan.