search this blog

Tuesday, April 21, 2015

West Eurasian mtDNA lineages in India


It might be interesting to compare the modern-day Indian mtDNA from this paper to the the ancient European mtDNA from Haak et al. 2015. The data is freely available in the spreadsheet here. Anybody up for it?

There is no indication from the previous mtDNA studies that west Eurasian-specific subclades have evolved within India and played a role in the spread of languages and the origins of the caste system. To address these issues, we have screened 14,198 individuals (4208 from this study) and analyzed 112 mitogenomes (41 new sequences) to trace west Eurasian maternal ancestry. This has led to the identification of two autochthonous subhaplogroups-HV14a1 and U1a1a4, which are likely to have originated in the Dravidian-speaking populations approximately 10.5-17.9 thousand years ago (kya). The carriers of these maternal lineages might have settled in South India during the time of the spread of the Dravidian language. In addition to this, we have identified several subsets of autochthonous U7 lineages, including U7a1, U7a2b, U7a3, U7a6, U7a7, and U7c, which seem to have originated particularly in the higher-ranked caste populations in relatively recent times (2.6-8.0 kya with an average of 5.7 kya). These lineages have provided crucial clues to the differentiation of the caste system that has occurred during the recent past and possibly, this might have been influenced by the Indo-Aryan migration. The remaining west Eurasian lineages observed in the higher-ranked caste groups, like the Brahmins, were found to cluster with populations who possibly arrived from west Asia during more recent times.

Palanichamy et al., West Eurasian mtDNA lineages in India: an insight into the spread of the Dravidian language and the origins of the caste system, Human Genetics, 2015 Apr 2. [Epub ahead of print]

263 comments:

1 – 200 of 263   Newer›   Newest»
Krefter said...

It should take several weeks to a few months at most to suck out the info from the supp. info. A comparison of the Indian WE mitogenomes to Haak 2015 mitogenomes will be very interesting. That will take a day or so.

Supp. info only gives the HV1+HV2 results of just under 1,200 samples in West Eurasian haplogroups. 11% of the 14198 samples fall under WE hgs. so there should be close to 1600. Maybe there's multiple samples in some haplotypes.

We can get an accurate idea how popular each subclade-haplotype is in the various Indian pops(geography, religion, etc.).

The 112 Indian WE mitogenomes are given in Supp. on a tree which is probably the best part. It would take a day to suck out the info from that. I can't figure out how to find where in India the mitogenomes come from though.

Kristiina said...

If U7 is connected with Indo-Aryan migrations, this female migration did not come from Europe as U7 is a typical Near Eastern and Indian haplogroup.

According to “Complete Mitochondrial DNA Diversity in Iranians” (http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0080673), U7a is the most diverse clade of U7 and it is found in virtually every population where the U7 haplogroup has been identified and sampled. The phylogeny of U7a reveals at least five sub-clades (with similar coalescence time estimates varying from 9 to 20 kya), which are most frequent in India, Iran and the Near East. Interestingly, a large subset of Iranian U7a samples could not be ascribed to any of its known sub-clades thus showing that more sampling of Iranian and Indian populations is highly desirable to uncover the ample diversity of this haplogroup. The second main sub-clade, U7b, dates to ~6–11 kya and includes several mtDNAs found predominantly in Europe, indicating a likely evolution in situ.”

In contrast to U7a, U7b seems not to be typical in India, but U7 may have originated in Iran and Pakistan area. It is nothing new that mtDNA and yDNA may take different paths, so Indo-Aryan maternal and paternal haplogroups need not have the same geographic origin although it is not excluded either.

Davidski said...

There are some specifically Asian mtDNA lineages in the Andronovo Kurgan remains, so it seems like the early Indo-Iranians weren't strangers to picking up women as they moved east.

Krefter said...

I'm surprised to see a mere 10% or so on average of mtDNA in India is "West Eurasian". There's another study I've seen with data from Pakistan and it's consistent with no regions having over 50% WE mtDNA. The WE mtDNA from India and Pakistan I've seen so far looks more West Asian-related than European-related. But I could be wrong.

bellbeakerblogger said...

@Kristiina

Most of the other maternal lineages seem to come from Northern Iran or Iraq as well.

It's also interesting to note those lineages that are absent.

bellbeakerblogger said...

I forgot to mention this.

DNA samples have been taken from four Harrappan individuals in a small plot being excavated now. They include two men, a child and a woman.

I'll take a wild guess that Mehrgarh II to Early Harrappan city dwellers had a substantial influx of R2 paternal lineages from the Middle East, which flourished and later spread with pre-historic archaeological movements along the Ganges River.

Early Western IVC mtdna may look very similar to those described in this study.

Nirjhar007 said...

David,
//There are some specifically Asian mtDNA lineages in the Andronovo Kurgan remains, so it seems like the early Indo-Iranians weren't strangers to picking up women as they moved east.//
Well BMAC people (Indo-Iranian) influenced Andronovo (Pre-Scythian people) etc so we can expect surely southern influence.
BB,

How come R2 Middle Eastern?O_o its surely SC Asian-Central Asian.

//Most of the other maternal lineages seem to come from Northern Iran or Iraq as well.//
Yes Quite right there is nothing much ''European'' either in Mtdnas or in Y-DNAs at the Subcontinent you guys will see in couple of months directly....

DMXX said...

Davidski,

"There are some specifically Asian mtDNA lineages in the Andronovo Kurgan remains, so it seems like the early Indo-Iranians weren't strangers to picking up women as they moved east."

I'm in full agreement.

Before anyone invokes the oft-used "misogyny" argument against Davidski's point, familiarise yourselves with the numerous reconstructions of proto-Indo-European culture from various sources (pantheon, kinship terms etc.). All clearly indicate the PIE's were a patriarchal group. The presence of female warriors in steppe (Amazons/European Scythians) and settled (presumably Median tomb) Iranians is irrelevant.

This, in turn, doesn't automatically assert an intentionally male-mediated demic movement in prehistory. Rather, new female lines were easily picked up in comparison to male ones.

As such, I do suspect we will observe mtDNA overlap between Andronovo, Sintashta etc. and other steppe groups with the Indian Subcontinent, albeit to a lesser extent than the Y contribution.

Another excellent example of this is the Xiaohe cemetery nestled between Gansu province and the Tarim basin in China; all of the males carried Y-DNA R1a1a (West Eurasian) whereas the maternal lines were predominantly mtDNA C4 (East Eurasian). The admixture event in my opinion likely took place in Afanasievo (widely held as proto-Tocharian).

DMXX said...

Nirjhar007,

"Well BMAC people (Indo-Iranian) influenced Andronovo (Pre-Scythian people) etc so we can expect surely southern influence."

The BMAC folk were clearly originally a non-Indo-European agricultural phenomenon. Their immediate origins are firmly secured as coming from the South Caucasus-North Iran area well before the earliest reconstructions of PIE (sometimes referred to as "Indo-Hittite").

The root BMAC culture had all sorts of cultural motifs that cannot be reconstructed in PIE. As such, they cannot be the chief linguistic benefactors of Proto-Indo-Iranian. For all intents and purposes, they're best described as an aboriginal agricultural non-IE phenomena in S-C Asia that interacted heavily with the early Indo-Iranians.

There are some rather strange patterns beyond S-C Asia which may be linked in some way with the BMAC. The "hot problem" in Tocharian studies right now is the absence of agricultural products in Afanasievo, but presence of these in proto-Tocharian. Additionally, one will observe Y-DNA J (varying percentages) in places like Gansu and Mongolia. There's no documentation of BMAC settlements east of Tajikistan.

Interesting as they are, they simply don't model well as direct derivatives of the Indo-Iranians for numerous reasons.

Krefter said...

@Niri,
"Yes Quite right there is nothing much ''European'' either in Mtdnas or in Y-DNAs at the Subcontinent you guys will see in couple of months directly...."

Do you know something we don't. Has someone given you a leak on unpublished ancient DNA results?

DMXX said...

^ Judging by the sentence content Krefter, I'd presume he's referring to some sort of upcoming paper on modern Subcontinental uniparental markers.

At this point in time, it is abundantly clear the majority of Indian Y-DNA R1a1a is Z93. The evidence continues to mount that Z93 arose somewhere near/in the West Eurasian steppelands (lgmayka at Anthrogenica posted about confirmed East European Z93* before I took my administrative hiatus). All current indications are that Z93 is itself a recent (<5000ybp) offshoot and spread into Asia through the Indo-Europeans (Indo-Iranians specifically it appears).

Whether or not R1a-Z93 may be classed as "European" is debatable* for numerous reasons (geographic dominance in Asia and not Europe, unknown area of origin in a transitional nexus between Europe and Asia etc.). Regardless, it is abundantly clear that between 15-20% of Indian (nationwide) Y-DNA lines (R1a1a) originate from outside of the Subcontinent in a region lying to the West at a very recent juncture.

I haven't followed up on sahab Nirjhar's posts to know his position and won't presume anything, but as a side point, the Out of India theory for anything pertaining to Y-DNA R1a1a, IE's or Indo-Iranians is (outside of India) clearly a defunct position given our current information and only a miraculous set of archaeological, linguistic and genetic counter-evidence are in order to affirm anything that runs contrary to the Eurasian steppe theory.

* I consider both a European and Asian classification as wholly meaningless. Prefer "West Eurasian" myself personally as it agrees with the similar ambiguity we have concerning exactly where Z93 arose, for now.

Colin Welling said...

@DMXX

Before anyone invokes the oft-used "misogyny" argument against Davidski's point

Sexism does not equal misogyny. Women aren't always being disadvantaged or being disadvantaged to a greater extent than men when it comes to sexism.

In many ways societies are harsh to men.

Seinundzeit said...

I've always been struck by how South Eurasian mtDNA lineages enjoy a steep peak after one crosses the Indus river. Pashtuns live west of the river, and are a transitional population in relation to West, South, and Central Asia. Looking at 90 Afghan Pashtun samples, 80% of the mtDNA lineages are West Eurasian, 13% are South Eurasian, and 7% are North Asian. Looking at 19 HGDP Pashtun samples, 85% of mtDNA lineages are West Eurasian, 15% are South Eurasian. These are consistent findings, Pashtun mtDNA frequencies tend to breakdown as around 80% West Eurasian, and around 10%-20% South Eurasian.

The Kalash also live west of the Indus, and 100% of their mtDNA gene pool is West Eurasian. Also in the neighborhood, the Burusho are only 10% South Eurasian in terms of mtDNA.

But when we look at neighboring Punjabis and Sindhis, the picture is quite different. In the Punjab, 50% of the mtDNA gene pool is West Eurasian, while 50% is South Eurasian. Although, one study had Punjabis at 40% West Eurasian mtDNA. Same with Sindhis, around 50% West Eurasian mtDNA. That's quite a gap, from 80% among Pashtuns to 50% among northwestern South Asians. Yet, the gap in autosomal genetics is significantly smaller.

For what it's worth, the largest mtDNA haplogroup among the 19 HGDP Pashtuns is W3a1, followed by different kinds of mtdna U and J. Among the 90 Afghan Pashtuns, mtDNA U is dominant, followed by different kinds of H.

Various sub-lineages of mtDNA W are quite common in Punjab as well. If I'm not mistaken, Yamnaya also had mtDNA W?

DMXX said...

Colin,

That is indeed the case. Without wading into semantics, some previous posters here appeared to display a zealous desire to attribute any hypotheticals involving male-mediated gene flow as an inherent bias of some sort. Such attempts (aside from serendipitously supporting a self-serving allegory) are completely irrelevant in discussions pertaining to different prehistoric cultures.

In the example used above, several fairly intuitive conclusions inferred from reconstructed PIE (prehistoric) culture point towards patriarchy. That was the specific point I was emphasising here in support of Davidski's post.

Davidski said...

From the paper...

"In India, only the eastern European-specific U5a1-derived lineages (Richards et al. 2000) such as U5a1a1, U5a1b, U5a1b1, and U5a1b1f were observed mostly in populations from Andhra Pradesh, Tamil Nadu, Uttar Pradesh and Punjab (Supplementary Table S1)."

"Indian U2e1a1 and U2e1b mitogenomes cluster with central Asian and Eastern Europeans (Palanichamy et al. 2004; Derenko et al. 2013, 2014) (supplementary fig. 3)."

This is enough to prove a steppe origin for a part of the Indian mtDNA gene pool. We can expect that the ratio of male lineages from the steppe was much greater, due to patrilineality and founder effects, and it does seem that way; note the high incidence of R1a-Z93 in South Asia today.

Mike Thomas said...

@ DMXX

Not saying you're wrong , but statements like "The BMAC folk were clearly originally a non-Indo-European agricultural phenomenon"
suggest you might be conflating (admittedly favoured) hypothesis with fact.

Some might accuse such statements of bias...

DMXX said...

Mike,

I would respectfully advise those readers to read into the archaeological background of the BMAC - as reported in literature, both recent and old.

Assuming sufficient comprehension and impartiality on their parts, they will report the same content or paraphrase similarly.

Given the prehistoric nature of the discussion, we're certainly dealing with degrees of likelihood. However, based on all available information, the likelihood that the BMAC was "Indo-Iranian" (in contradistinction with Andronovo being "pre-Scythian" as Nirj stated) is just not very likely at all.

The material culture of the BMAC make it abundantly clear that, for them to be originally Indo-European, a most remarkable act of archaeological deception and cultural camouflage is in order. There is a palpable mismatch between the reconstructed PIE vocabulary, deities and environment with the BMAC. Andronovo fits the bill much better (although not cleanly).

Mike Thomas said...

perhaps, but that palpable incongruity beyween BMAC and "reconstructed PIE world" is only an issue if one treats 'Porto-lexical" exercises relevant for anything over and above chronology; and can rest assured that they in the first place were constructed free of a biased framework which didn't "work backwards" and post hoc

The answer to both is negative imho

Now, I ask you for the evidnece of Andronovo or other "indoeuropean" intrusion into BMAC. ....

Nirjhar007 said...

WOW DMXX Welcome! I Was checking your blog 2 days ago quite nice why did you stop posting? your post on N Y-DNA in India is quite impressive.
//Interesting as they are, they simply don't model well as direct derivatives of the Indo-Iranians for numerous reasons.//
Well dude you are mistaken i didn't say BMAC as IIR originator just that they spoke Indo-Iranian and influenced Sintastha,Andronovo etc with their Civilization possibly bringing Indo-Iranian influence on Uralic.
Your conception on ''Indo-Iranian'' is old and out dated! doesn't matter how famous the notion is AND you can get a general idea on the Whereabouts of Indo-Iranian people here-
http://new-indology.blogspot.in/2013/02/indo-iranians-new-perspectives.html

Nirjhar007 said...

Mike, Linguistic arguments supporting non-IE nature of BMAC is as good as photographic arguments for UFO's....

DMXX said...

Mike,

A simple search online for "Andronovo" and "BMAC" will yield all the answers you're looking for. Off-hand, there's some very good examples of Andronovo-BMAC hybrid sites found in western Tajikistan. Alternatively, Kuz'mina's The Origin of the Indo-Iranians is the most exhaustive current body of evidence we have. If you can't find it online, please get in touch. Would be glad to help.

Nirjhar,

Thanks for the kind comments, very much appreciated! Time is unfortunately a finite quantity and I've been forced to prioritise my online attention on co-running Anthrogenica. Having said that, I have several entries waiting to be finished off which the community might find interesting.

What literature has convinced you the overall interaction was south-to-north and not the other way round? If I recall Kuz'mina's The Origins of the Indo-Iranians, we see evidence of Andronovoid groups initially settling on the outskirts of BMAC, exchanging material goods before eventually culturally usurping the territory and absorbing much of the previous culture. Including their language and certain religo-cultural motifs.

Forgive me for summarising this part in such a shorthand way. The book is hundreds of pages long. I can refer you to precise pages and quotes at a later time if need be.

Nirjhar007 said...

DMXX,
//I have several entries waiting to be finished off which the community might find interesting.//
Thats great! i just can't wait.
//What literature has convinced you the overall interaction was south-to-north and not the other way round?//
Its just not ancient literature but by putting Rationality and Some Archaeology and Anthropology also and yes major influences went south to north the opposite of that is in one word ''insignificant'' or ''dim'', i know that book quite well BTW read the article also when you get time:).....
Cheers...

DMXX said...

Nirjhar,

Duly noted. I've bookmarked the article you shared and will give it a read in several week's time (my hiatus everywhere is study related). I tend to avoid wholesale subscription to any particular viewpoint and follow the data, so I do look forward to seeing what's being proposed here. Thanks for having this resourceful discussion with me.

Nirjhar007 said...

DMXX,
Jolly Good:).

Krefter said...

This is a spreadsheet with the Indian full-sequence samples in WE haplogroups from the tree in Supp. 2. There are well under 100, and I don't know where the other 40 or so are.

Supp. 2. challenges some of the classifications of Phylo, and there are some interesting trends in the non-Indian samples.

I can see H13a1a, H2b, U5a1b1, U2e1, U4a1, W6, W3a1, I1, and R1a may be significant shared clades between LN/BA, Yamna, and India. Most probably are just from shared Near eastern and Ancient Western Eurasian maternal ancestry though. I haven't looked at the data thoroughly yet.

https://docs.google.com/spreadsheets/d/10l6uVHsH-ezT6a9VlomE8jDkR9LjD9K3A4lFcdUOe-k/edit#gid=0

Here are the full-sequence samples from Haak 2015.

https://docs.google.com/spreadsheets/d/1exPddmIQWeCu3wdBY9XpRZ5NyD2lkj5MuvhfJdyxWwk/edit?usp=sharing

And here's the Ancient mtDNA package.

https://drive.google.com/open?id=0B3_CjQDOYA7GVVRlcW5oZnpINzg&authuser=0

Mike Thomas said...

DMXX

Whilst my question show me the evidence" was a rhetorical one, I appreciated your answer nevertheless.

I am familiar with Kuzmina's works - a great scholar indeed. To me, her 'acculturation and usurpation' theory only highlights the lack of evidence of a of steppic intrusion into the area of BMAC, and ultimately must fall back onto colorful theories and witty metaphors. I guess, to be impressed, I'd like to see a rather more convincing modelling of the archaeological evidence.

There is no evidence of destruction dating to any concise period which might coincide with a conquest. Nor is there any evidence of a 'usurpation'. Nor is there any explanation as to how a few pastoralists traders bringing in milk and cheese managed to pull the 'wool over the eyes' (pun intended) of the local elite. Moreover, these elite operated as a heterarchy most likely, making any coup all the more difficult - necessitating the whole scale uprooting of an entire elitist merchant class rather than replacing one king.

So, at the moment at least, the theories of 'acculturation' of yesteryear are not convincing to me, and I rather look more sharply at the evidence - which shows occasional trade between mobile pastoralists and the elite of the Darya region limited to regions at the periphery of BMAC - seen by the milky residues examined in the 'barbarian steppe' ceramics found, and for a long time (Wrongly) seen as evidence of acculturation' - crude analysis lacking any contextual consideration.

aDNA might prove an invasion scenario - but if so, and if north to south- then it didn;t happen in BMAC region

Mike Thomas said...

Krefter

Thanks for you mtDNA breakdown.
Is it just me, but why aren't the maps links working ?

Balaji said...

mtDNA Haplogroups

Even by 1999, it was known that the Indian Subcontinent and Europe shared common ancestry dating back to the late Pleistocene. Kivisild et al “Deep common ancestry of Indian and western-Eurasian mitochondrial DNA lineages”.

http://www.sciencedirect.com/science/article/pii/S0960982200800573

In that article, the authors identified U2e as Europe-specific and U2i as India-specific. In the years since, U2i has split into U2a, U2b and U2c, all India-specific. Moreover, U2e is now known to be present in India, the earlier lack was due to sampling bias. U2 must have originated in India and U2e spread from there to Central Asia and Europe. MA1 is of haplogroup U2e.

In a 2004 paper, Metspalu et al. showed that mtDNA haplogroups U7, W and R2 are West Eurasian lineages that hove been in the Subcontinent for tens of thousands of years.

http://www.biomedcentral.com/1471-2156/5/26


Y Haplogroup R

Y-haplogroup R is analogous to mtDNA haplogroup U2. R2 is India-specific but R1 is found both in the Subcontinent and outside it. See the following table which has frequencies of different Y-haplogroups in various Indian populations, both high-caste and low-caste.

http://www.nature.com/jhg/journal/v59/n7/fig_tab/jhg201442t1.html#figure-title

A few instances of R* were found. But R1* was not so rare and several populations had it. R1a* was also found in two populations with a tribal population having 23% R1a*. All this suggests that R originated in the Subcontinent and also differentiated into R1 and R2 within the Subcontinent. MA1 has Y haplogroup R*.

Haak ADMIXTURE figure for MA1

http://biorxiv.org/content/biorxiv/suppl/2015/02/10/013433.DC1/013433-1.pdf

MA1 shows East Asian yellow for K=3 through K=5. But at K=6, this is completely replaced by Papuan purple. At K=8, Papuan purple is completely replaced by South Asian lime green. At K=11, an Onge beige appears and MA1 has it. MA1 then retains the South Asian lime green and the Onge beige for all subsequent K values.

The other two Pleistocene-era ancients, Ust Ishim and Kostenki14 are “rainbow coalitions” at all K values showing affinities to virtually all modern humans. This is because they are so early that they are before the time that humans from different areas had diverged much. MA1, however, is younger and does not show East Asian yellow after K=5 and Papuan purple after K=7. However, MA1 has South Asian lime green and Onge beige at all K values after they first appear. This indicates that the ancestors of MA1 originated in the Indian Subcontient at a time when ANE had not yet differentiated too much from the Onge-like ASI.

D Statistics

Based on what I see from the Haak ADMIXTURE figure, I predict that D(WHG, MA1; Yoruba, Papuan) > D(WHG, Samara_HG; Yoruba, Papuan) > D(WHG, Karelia_HG; Yoruba, Papuan) > D(WHG, Motala_HG; Yoruba, Papuan) > 0.

Davidski, can we see if it is so?

Krefter said...

@Mike,

I have no idea why you and others can't see my maps with ancient mtDNA. Sometimes when I post Google drive files some people can't even see the Documents and spreadsheets.

Maybe it's because they have a Mac or don't have a Google account, I don't know.

Davidski said...

Balaji,

MA1 doesn't belong to U2e. It belongs to U*. Kostenki14 belongs to U2.

Also, these haplogroups aren't just shared by Europeans and Indians, they actually come from Eastern Europe.

U5a1a1
U5a1b
U5a1b1
U5a1b1f

There's no other way to put this.

Krefter said...

Balaji,

Thanks for the links to those papers. I agree U2e is not foreign to South Asia or European-specific. As far as I know it is widespread all over West Eurasia, South Asia, and Siberia. Thanks to Ancient DNA we also know it was in Germany 11,000YBP and Sweden and Russia 8,000YBP.

The U2 clades which are popular in South Asia also exist in West Asia. I think we need full-sequences of U2 from all over Eurasia to get an idea where each clade is most diverse(doesn't mean it originated there). It is very important to remember 38,000YBP K-14 had U2-his own basal clade(which is now extinct), confirming U2 is very old.


"A few instances of R* were found. But R1* was not so rare and several populations had it. R1a* was also found in two populations with a tribal population having 23% R1a*."

I've read studies with R*, R1*, R1a*, and R1b* in west and south Asia. R1a1* and R1b1* in Mesolithic Russia doesn't make these disappear. Something is happening and it needs to be investigated. Although I think the most popular modern R1 clades: R1a-M417 and R1b-L23 are probably of East European origin.

"This indicates that the ancestors of MA1 originated in the Indian Subcontient at a time when ANE"

Saying MA-1 came from South Asia is a bit of a stretch. His closest relatives are EHG, Native Americans, and then WHG. WHG/UHG existed in Europe and west Asia.

Krefter said...

"U5a1a1
U5a1b
U5a1b1
U5a1b1f

There's no other way to put this."

Two of the Yamna guys had U5a1a1.

Mike Thomas said...

Balaji

Thanks for the breakdown.

Yes, Hammer et al (2001) found that there was a "predominance of intercontinental sig-
nals detected emanating from Asia (Æg. 4). These multiple out-of-Asia signals included gene flow episodes to Europe and the Americas, along with range expansions
to Oceania, Africa, and Europe. In contrast, the NCA only detected two out-of-Africa signals."

So to me it looks obvious that R, and maybe even R1 arose in South-central Asia. But the more proximate history of the M417 dervived groups might be a different story (?)

Moreover, Mal'ta might be a failed/ abortive colonisation of Siberia

Nirjhar007 said...

Mike,
//But the more proximate history of the M417 dervived groups might be a different story//
Just need the aDNAs coming out of SC Asian sites to clear it all...

Nirjhar007 said...

David,
//There's no other way to put this//
Whats the age of those clades?

Nirjhar007 said...

Balaji,
IE Upper castes have typically
H and U7 and W, which are well present also in Iran. Unfortunately, it would be interesting to see
also which H... Curiously, Middle Caste Dravidians have more Western Hgs than Upper caste, do you
know what is Middle Caste there?

capra internetensis said...

@Balaji

It's really hard to say where R originated, though South Asia is certainly a prime suspect.

I wouldn't call R2 "India-specific", though R2a-M124 is certainly most frequent there. The older studies did not test for R2-M479, so all reported R2 is now R2a-M124, and R*(xR1) can actually be R2*. A couple of recent studies did check for M479, and Di Cristofaro found that the R* in Pakistan and Afghanistan - very common in Burushos - is all R2*. In fact only one true R*(xR1, R2) turned up in that study - a Kyrgyz.

Low levels of R2a are found very widely, and you can find R2* in Iran and Central Asia, peaking in Pakistan and Afghanistan. We really need to know more about the structure of R2 to be able to tell whether it comes from India, or South Central Asia, or elsewhere.

When it comes to R1, we really need some up-to-date and high resolution studies of India. There is that new branch of R1b* in Tajikistan and Bhutan which is almost certainly in India as well. Underhill found that Indian R1a was almost entirely Z93, but older studies have found a little bit of R1a1*-SRY1038.2 and also basal R1* which could be R1a*-M420. In fact the study you linked found a surprising amount of R1*(xR1a1, R1b) and I have seen nothing in subsequent studies giving the slightest hint what it actually is.

------------------------

I looked for some suitable Onge D stats, not quite what you are looking for but maybe helpful:

Loschbour, Chimp : Ami, Onge 0.000511 1.255
MA1, Chimp : Ami, Onge 0.000578 1.402
Onge, Chimp : LaBrana, Stuttgart 0.001528 2.797
Onge, Chimp : Loschbour, Stuttgart 0.00191 3.452
Onge, Chimp : MA1, Stuttgart 0.001842 2.987
Onge, Chimp : Motala12, Stuttgart 0.002105 3.66

MA-1, WHG, and Motala12 don't differ significantly in affinity to Onge using these outgroups.

Krefter said...

Thanks for the info capra.

"There is that new branch of R1b* in Tajikistan and Bhutan which is almost certainly in India as well."

Do you mean it was negative for all known basal branches of R1b? That would exciting news.

BTW, AG2 an individual very similar to MA-1, 10,000 years younger, and from the same general location as MA-1 is P1+ and R1+. He may have been R1.

Tobus said...

Balaji:
Loschbour MA1 Yoruba Papuan 0.0122 2.207 11882 11596 251027
Loschbour Samara_HG Yoruba Papuan 0.0085 1.651 9184 9029 204902
Loschbour Karelia_HG Yoruba Papuan 0.0011 0.229 15097 15063 338485
Loschbour Motala_HG Yoruba Papuan 0.0012 0.322 14264 14229 343080

Tobus said...

.. and for completeness:

LaBrana1 MA1 Yoruba Papuan 0.0079 1.399 11250 11075 239972
LaBrana1 Samara_HG Yoruba Papuan 0.0038 0.697 8794 8727 197044
LaBrana1 Karelia_HG Yoruba Papuan -0.0025 -0.494 14435 14507 323706
LaBrana1 Motala_HG Yoruba Papuan -0.0036 -0.901 13947 14048 327973

HungaryGamba_HG MA1 Yoruba Papuan 0.0098 1.686 8216 8057 174887
HungaryGamba_HG Samara_HG Yoruba Papuan 0.0116 1.923 6357 6211 143566
HungaryGamba_HG Karelia_HG Yoruba Papuan 0.0037 0.657 10452 10376 235365
HungaryGamba_HG Motala_HG Yoruba Papuan 0.0044 1.055 9976 9888 238490

Nirjhar007 said...

Capra,David,Balaji,Mike.
U5 is typical of Hunter Gatherers as you know, so it can be arrived even in Paleolithic!
It is really remarkable that U Hgs not-U7 (which is Iranian) are most frequent in the South,
so it seems that Hunter Gatherers Hgs are more preserved there?.
And also it speaks of Andhra and Tamil Nadu.

Davidski said...

The Eastern European-specific subclades of U5 did not arrive in India during the Paleolithic. They didn't exist at that time.

Nirjhar007 said...

David,Where i can find a good estimate of these clades ages?
U5a1a1
U5a1b
U5a1b1
U5a1b1f

Davidski said...

Here...

http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0010285

U5a1a1 is only around 12,000 years old and U5a1b only around 11,000 years old. U5a1b1 and U5a1b1f are obviously younger. These might be overestimates anyway, as is often the way.

There weren't any population movements between Europe and India at that time, and these haplogroups obviously have nothing to do with early Neolithic farmers.

The only explanation is gene flow from Europe to India with Late Neolithic/Early Bronze Age groups via the steppe.

Nirjhar007 said...

Common idea is U5a1a1 developed 12 KYA.
U5a1b 8 KYA.
U5a1b1 7 KYA.
Interesting....

Nirjhar007 said...

Thx!,
//There weren't any population movements between Europe and India at that time, and these haplogroups obviously have nothing to do with early Neolithic farmers.//
It appears so.
//The only explanation is gene flow from Europe to India with Late Neolithic/Early Bronze Age groups via the steppe.//
It have to be 4500-400 BC otherwise no chance Davo....
As In case of Mtdna Intrusion i think the migrations have to be more substantial...

Mike Thomas said...

Dave

What is the best way to analyse affinities of the ancient samples to modern pops?
Can you do a run for the three Uneticians from Haak?

Krefter said...

@Davidski,

Those U5a1s in India could have really complex origins. Just about every clade is found everywhere it seems. There's even U5b1b1a in the Near east. There are lots of differnt ways those U5a1 clades got to India.

Mike Thomas said...

Also Dave

Im not sure about some of the conclusions of that Malyarchuk paper:
Ie the supposed doscontinuity in EE during the LGM. Fact is, there might have been greater continuity in EE than Central Europe. And we have U5a from Mesolithic russia, Ukraine etc, whilst U5b was in Iberia.

We also know that U5b was in the Balkans (Mesolithic Korcula) and U5a in Minoans (? a proxy for pre-Neolithic Southern Balkans)
Details are still sketchy, however.

Nirjhar007 said...

Krefter, U5 Mtdna like R2 Y-DNA in case of India is Exclusively European looking so the simple Explanation is that U5 and their Subclades spread from Europe or surrounds?
Though It will be interesting to see if any Ancient Asian Sites yields U5 s or not.

capra internetensis said...

@Krefter

Yes, they form a clade together which is basal to all other known R1b, including V88.

From what I recall AG2 was so badly contaminated that his Y hg could not be determined (modern DNA from Russian archaeologists will get you R1 positive for sure).

@Balaji

Your prediction seems to be mostly correct, but there is some weirdness here. For two of the references Motala_HG is more Papuan than Karelia_HG, even though the latter has more than double the ANE, and some East Asian on top of that. Samara_HG is closer to Papuan than MA-1 is when HungaryGamba_HG is the reference. And La Brana is more Papuan than either Motala_HG or Karelia_HG!

The problem with Papuan is that it has a lot of Denisovan, and I think that might be picking up differences in archaic admixture. Or it could just all be noise from low-quality ancient samples, I guess.

Davidski said...

Nirjhar,

The fact that Indians carry a wide range of relatively young Eastern European-specific mtDNA lineages means that there were recent population movements from Eastern Europe that eventually affected India.

There's no threshold for this. They're in India, so it happened.

Krefter,

We know the detailed phylogeography of these young U5 lineages, based on both modern and ancient samples. They look like they expanded from Eastern Europe, and not from the Middle East or India.

Mike,

Just have a look at the IBS stats and where the Unetice samples fall on the PCA of West Eurasia done by Matt and myself.

capra internetensis said...

My guess is that ANE is really more Ancient Central Eurasian, so to speak, extending probably down to northern South Asia and eastern Iran even in the Paleolithic; I doubt it has any special affinity to ASI in itself, but ANE coming from the southern part of its range would have a little ASI admixture. I guess that Samara_HG has a bit of that, while Karelia_HG has a bit of northeastern, East Asian-mixed ANE. But I could be completely wrong.

Mike Thomas said...

Ive been trying to goose around with PCAs

Here's one of ancient mtDNA data & different pops from Mesolithic Iberians to Bronze Age Kazakhs

https://drive.google.com/file/d/0B1vtTHobiXwVVDM4VXlaUWpZTjg/view?usp=sharing

I think it makes sense (?)

Davidski said...

Very nice. It makes sense to me.

Mike Thomas said...

MNG= mid Neolithic Germany
SWE HG = mesolithic Iberia (SW Europe)
MNSp= mid Neo Spain
HG CE= Mesolithic Central Europe

The rest are self- explanatory, I think

Nirjhar007 said...

David,
I understand what you are implying but Kurgan hypothesis which is Essentially wrong in Case of Asia and Indo-Iranians states the spread to be ''like of Turkic or Huns'' hence we don't see any archaeological trail that we can assume for example Europe with CWC GAC etc, a male dominated spread with absorbing Mtdnas certainly can't be put with Indo-iranians who show rather Rooted and vast presence in Asia if we bring rationality in the context and some observations and it contradicts the Steppe Nomads identity that is traditionally coined with them* and also projecting the spread with a Single Y-DNA subclade and to think that they came from E Europe which have bunk like Z93* distribution like we also see in case of R2* In Portugal etc with not leaving any detectable trails is ridiculous and , yes U5 subclades look like of European but again i don't think this can actually be put in the 2nd millennium bc context as that is a failed suggestion and they don't look that ''young'' and Mitochondrial Migrations if i'm not wrong Combined with Y-DNA also requires substantial population movements with Archaeological correspondence bringing new cultures that don't exist between E Europe and SC Asia so i think those lineages are quite old at least Chalcolithic and Even more And there of course is the vast want of aDNA data from Asia.....
*Scythian nomads were of Iranian Stock who are the Historical speakers of steppes they should be seen not as the bearers of Indo-Iranian languages from the north to the south, but the opposite, as the nomadic pioneers of the Iranian languages (like the Tuiryas and Sairimas of the Avesta), who brought them up to Siberia in the east and Ukraine in the west.

Mike Thomas said...

Thanks Dave

Seems there truly is a Copper-Bronze age continuum from Central Europe to Northern Central Asia.

Interesting that there was some kind of PalaeoSiberian shift in the catacomb period. Perhaps the old skool archaeologists were right for this one .

I forgot to add Minoan data. I'll do that tomorrow

Alberto said...

@DMXX

"only a miraculous set of archaeological, linguistic and genetic counter-evidence are in order to affirm anything that runs contrary to the Eurasian steppe theory."

But does autosomal DNA, both ancient and modern, mean nothing?

We have ancient DNA proving a large scale migration from Central (?) Asia to the steppe. About 50% of the Yamnaya ancestry had this Pashtun-like signature.

In modern DNA we still see the clear influence of this admixture across Europe.

We have no ancient autosomal DNA proving or disproving any European migration to S-C Asia. But modern DNA shows roughly 0% European admixture in these areas.

I wonder what's your take on these facts. To simply ignore them?

Davidski said...

Alberto,

You're wrong when you say that there's no evidence of European admixture in Central Asia. There is such evidence both in genome-wide DNA (for instance, the blue cluster in my K=9 run) and Eastern European-specific U5 lineages.

Asian R1a-Z93 ultimately also comes from Eastern Europe. At some point this will have to be acknowledged by everyone here.

Tobus said...

Capra:Samara_HG is closer to Papuan than MA-1 is when HungaryGamba_HG is the reference. And La Brana is more Papuan than either Motala_HG or Karelia_HG!

The Z-scores on all of these except the Loschbour/MA1 one are insignificant, and even that one is borderline if you take the original |Z|>3 threshold. My take on these results is that they are essentially zero results and MA1, EHG, SHG and WHG all form a clade relative to Papuans - Papuans diverged before any the others and there hasn't been any admixture since. MA1 *might* be slightly closer, but then again he's ~15ky nearer the divergence point so that might be the reason for his increased scores.

Alberto said...

@David

"You're wrong when you say that there's no evidence of European admixture in Central Asia. There is such evidence both in genome-wide DNA (for instance, the blue cluster in my K=9 run)"

Yes, the blue does show in your K9. But it seems related to ANE. I think it's a similar issue as the "West Asian", in the way that Chad explained the other day quite nicely. When using the European specific WHG (in K8) it shows basically 0%.

"Asian R1a-Z93 ultimately also comes from Eastern Europe. At some point this will have to be acknowledged by everyone here."

I was referring to autosomal DNA. I think it should count, not just be ignored. Re: R1a-Z93, it's still speculative. Maybe it will be proved to be of EE origin, maybe not.

But a large migration from Central Asia to the steppe (and the rest of Europe) is already proved. Apparently, they left a big impact in autosomal DNA but no trace of uniparental markers.

OTOH, Europeans went to S-C Asia and left no trace of autosomal DNA, but a big impact on uniparental markers.

Kind of strange, if you ask me. But time will tell.

Davidski said...

Alberto,

There blue cluster reflects gene flow from Europe, because it shows a good correlation with European-specific mtDNA lineages in Central Asia.

So your comment that there's no evidence of European autosomal admixture in Central Asia is not accurate. At the very least there are strong hints that this is indeed the case.

In regards to R1a-Z93, there's very little doubt that it's a steppe marker closely related to the Balto-Slavic R1a-Z282, and also the ancient Karelian R1a-M459.

I do understand that proving this without any doubt will basically clinch the PIE steppe hypothesis, so I'm expecting we'll need a lot of ancient DNA before everyone accepts it. But we'll get to that point eventually.

postneo said...

@dmxx

"The material culture of the BMAC make it abundantly clear that, for them to be originally Indo-European, a most remarkable act of archaeological deception and cultural camouflage is in order. There is a palpable mismatch between the reconstructed PIE vocabulary, deities and environment with the BMAC. Andronovo fits the bill much better (although not cleanly)."

It's bullshit. There is no match or mismatch with either andro novo or bmac. First of all there is no reconstructable unified monolithic PIE culture. You can easily pick up isolated resemblances as sarianidi and others do and theorize.

Chimps(non bonobos) are mainly patrilineal and patrilocal .. so what.

The only genuine match with statistical significance are Mitanni ...Its hard to fake the sequence of syllables., hittites and early greeks.

Mike Thomas said...

@ PostNeo

"The only genuine match with statistical significance are Mitanni ...Its hard to fake the sequence of syllables., hittites and early greeks."

What do you mean exactly ?

Alberto said...

@David

I don't know why suddenly you trust more those admixture runs with mixed components than your own more scientific ones.

There might be some European admixture in S-C Asia, but if there is it's pretty small. You know better than anyone that the gene flow is clearly in the opposite direction. All I'm saying is that it cannot be ignored, as the above poster seemed to be doing.

Nirjhar007 said...

David,
//I do understand that proving this without any doubt will basically clinch the PIE steppe hypothesis, so I'm expecting we'll need a lot of ancient DNA before everyone accepts it. But we'll get to that point eventually.//
We Just need the Indian aDNA actually to settle the Z-93 and M-417 Issue its that simple...
About PIE well IMHO A single Y-DNA clade is not very decisive.
Postneo
//The only genuine match with statistical significance are Mitanni ...Its hard to fake the sequence of syllables., hittites and early greeks.//
You mean ancient written texts should taken as the basis to evaluate PIE Origins?

postneo said...

@Mike @Nir

yes I meant written texts are the only genuine evidence so far.
so you have linear-B, luwian, hittite, mitanni etc.

Nirjhar007 said...

David, Mike, Alberto and others
I was discussing the case of U5 somewhere else and let me tell that I don't trust much genetic dating, as usual, U5 is particularly present in Mesolithic sites, so maybe
it arrived already in the Mesolithic, but it could arrive also later through Iran and Central Asia.
According to "Complete Mitochondrial DNA Diversity in Iranians" (by Derenko), in Iran there is a
specific subclade called U5a1a'g with the subclaster U5a1g, cited also here:
http://www.eupedia.com/europe/Haplogroup_U5_mtDNA.shtml#subclades
It is also found in Russia, Slovakia, England, and Italy.
in Genographic also in Romania, Jordan, Saudi Arabia, Greece, Ukraine and Germany... maybe it was
brought by IEs from Iran, but I don't know if there is in India!.

Chad Rohlfsen said...

I have a K10 that has a component peak in Yamnaya and separate from any central and south central Asians. I'm sending it to Favid later. I may upload it to google drive.

Chad Rohlfsen said...

To David, sorry. Fat fingered the phone.

Nirjhar007 said...

Chad,
You mean a Yamnaya specific component which is also present in C and SC Asians?.

aniasi said...

Most interesting. Nonetheless, the idea of such broad haplogroups being West vs South Eurasian is a bit ridiculous. Several South Asian clades of M are far older than U, and the oldest clades of N and R are found well rooted in the subcontinent.

Nonetheless, these 'West Eurasian' clades seem to sometimes concentrate amongst Dravidian speakers in the South, and their arrival seriously undermines the simplistic analysis and timeframe of ASI-ANI admixture.

These indicate migrations before the 4kya admixture signal, and more importantly, show that there is a differentiation between later and earlier West Eurasian clades.

What would have made this study more interesting is the inclusion of clades like M3 & M4, which show Northwestern Indian concentrations and may have been part of the same migrations to the South and East.

Nirjhar007 said...

aniasi,
//These indicate migrations before the 4kya admixture signal,//
Which admixture signal you talking about?.

aniasi said...

@Nirjhar,

If I am correct, the last major Indian population study indicated that there were two ASI-ANI admixture events that were signalled in South Asian DNA, one 4kya, and a later one that only happened in the North 1.6kya.

Nirjhar007 said...

But ANI-ASI ADMIXTURE Should be more old as People from West Asia Iran Came there in Chalcolithic period and also earlier and there were also Movements from Central Asia From Mesolithic.

aniasi said...

@Nirjhar007

As this new evidence shows. However no one has even begun to question or criticise the last study with this new information.

Nirjhar007 said...

Ah yes I can agree to that the Acceleration of ANI-ASI ADMIXTURE started around 4kybp without any doubt.

aniasi said...

@Nirjhar

From what I gather, though, the previous study made it out to be one large and quick event instead of continuous admixture.

However, how would admixture have accelerated after 4kybp? Surely that is when it stopped and South Indian castes were formed?

Nirjhar007 said...

Well that study was made by ROLLOFF to account the admixuture signals if i'm not wrong and it is not very powerful tool to detect archaic admixture process so i think the process was continuous and it had a bang around the 4.2 kilo year event when SSC Started to decline and the inhabitants were forced to settle in more Eastern and Southern Areas which IMO were ASI Dominant hence the acceleration, about caste my learning tells me its an archaic arrangement in IE Society and was there well before 2000 bc and as you know it got very strict around 2000 YBP....

capra internetensis said...

The study on ASI-ANI admixture modeled it as single pulses for methodological reasons, but the authors were careful to point out that this was not likely to reflect reality. In fact for most populations they said the pattern of linkage disequilibrium decay fit better with multiple or ongoing admixture episodes.

For a few populations, including the Vyshya of Andhra Pradesh who had the oldest admixture signal at 4400 years ago, the LD signal was consistent with a single episode, but even that could just have been limited resolution of the technique used.

The study is available for free online if you Google for it.

Nirjhar007 said...

Capra,I think if i'm not wrong that study had 2 previous versions with 2 different dates for ADMIXTURE Acceleration! probably 1500 BC and 2000 BC? before settling with 2200 BC Anyway i mostly agree with you.

Chad Rohlfsen said...

Yes, it's a pretty heavily Yamnaya like component.

Nirjhar007 said...

cool! will be very interesting to see.

Krefter said...

@Davidski,

Can you test rs86insA and Y152OCH, in my file and the Motala HG's and EHG's files? Interprone and 23andme can't find them.

jwray3 said...

David-
I'm sure it can be a bit burdensome responding to everyone's inquires, but one of these days when you have a break could you stop back over to Eurogenes Genetic Ancestry Project and the 'Eurogenes' GEDmatch ADMIXTURE ancestry test guide' post and take a look at a few of our recent questions.

Sorry for drifting off topic, you can delete this post

Chad Rohlfsen said...

Here are a few examples of what a few pops score in this component.

Yamnaya <65%
Corded Ware <53%
N Caucasus pops <49%
E Euro, Beaker, NW Euro <38%
S Caucasus pops <35%
Tajik <35%
Chuvash <35%
French <34%
SE Euro <32%
Pathan <31%
Levant <30%
Iranian <30%
Makrani, Balochi, Brahui <29%
Spanish <27%
Basque <26%
Uzbek <25%
Burusho, Sindhi, Turkmen <24%
Kalash, Gujarati, Uygur <23%
Krygyz <20%
Sardinian <16%
Punjabi <14%

Chad Rohlfsen said...

Siberians like the Evens and Yakuts score in the mid 20s as well.

DMXX said...

Mike,

I see you are well read. :) A couple of other publications have convinced me of the broad model proposed by Kuz'mina. I agree with your criticisms regarding her simplistic premise of apparently docile agriculturalists in the BMAC being swamped by an IIR elite. The linguistic evidence doesn't necessarily support this. I'll get in touch and provide you with some of the materials at a later date.

Alberto,

We must continuously caveat the discussion by acknowledging the absence of ancient DNA in the relevant regions discussed here.

I am in agreement with all of your post up to the final three sentences. Davidski has already addressed the genetic side of the argument. The modern auDNA clearly suggests the presence of a European-like signal deep into Asia. The distribution coincides too neatly with either steppe-based archaeological cultures or regions that do (or once) speak/spoke Indo-European languages. Even Dienekes' own Dodecad runs yielded this repeatedly.

As such, I have no idea what autosomal data you've been looking at. Clearly not what the rest of us have. Which I frankly find bizarre given this venue. Davidski's posted a lot about this here.

The genetic evidence aside, one must consider that the Indo-European enigma was always a linguistic problem. Other evidence (archaeological, genetic) always played an auxillary role. Gimbutas' original model was founded on archaeological grounds (paradoxically). Linguists have since continued to find support in the Eurasian steppe theory, to the point where it is (I've been told by two linguists) a near-consensus view these days after years of refining. This prevailing current theory appears to be increasingly fortified with the genetic side.

Hope that answered the thrust of your message. Also, try and practice some civility if you choose to reply to this message. I don't have the time or interest in playing agenda games (why I'd "simply ignore" genetic data is beyond me).

postneo,

"reconstructable unified monolithic PIE culture" (both in an archaeological and anthropological sense) isn't what is described here. We have no firm grasp regarding how the early PIE's functioned, merely reconstructed clues on things like kinship terms, rituals and their pantheon. On a more humorous note, I've always found it intriguing they had at least two words for flatulence... Given the Samara Yamnaya did not have lactase persistence, perhaps that's indication the pre-PIE's were drinking milk irrespective of what their bodies told them? :)

As for the rest of your message:

- Horses and horse domestication items were only observed in the BMAC after indications of Andronovo interactions took place. They weren't present beforehand. Conversely, we find the full line of evidence necessary on the steppelands. Horse remains, signs of horse consumption (especially in Botai) and finally, signs of horse domestication, are all found among steppe cultures. To make matters worse for opposing viewpoints on this topic, wheel/wagon terminology was shown by Anthony to become more extensive with each IE group that left the putative steppe homeland (http://kinshipstudies.org/wp-content/uploads/2015/02/KinshipStudies-IndoEuropeanWheelMap-copy-2.jpg). No coincidence the English word for "circle" (itself a loan from French IIRC) has cognates in Greek (kuklos), Persian (charx) and Hindi (chakra). There are no wheels in the BMAC (Anthony).

- Deities. One can write about this for hours. The BMAC folk had several animal cults which are not substantiated in any of the known IE lexicons (if one sets aside the likelihood these were loaned directly into Indo-Iranian). Specific example being the snake. I'll refer you to Kuz'mina's work for more details (Parpola's also published some convincing work on this IIRC).

Nirjhar007 said...

Chad,
Very Interesting Finally you have able to make a yamnaya component.

Davidski said...

Reading this recent archeological article it's not that difficult to see how the highly mobile steppe pastoralists from north of the Caspian might have picked up West Asian mtDNA lineages and Central Asian autosomal admixture. See Figure 1.

http://rspb.royalsocietypublishing.org/content/281/1783/20133382

But I'm still confused how the North Caucasus fits into all of this? At this stage the only thing I can think of is a massive migration from the steppe or Central Asia to the North Caucasus, perhaps with the speakers of proto-northeast Caucasian languages? The problem is that Georgians speak a Kartvelian language, and to me they look like the best proxy for the West/Central Asian admixture in Yamnaya.

Nirjhar007 said...

DMXX,
Hi since your suggestions are to Postneo i would like to let him answer but on this-
//Deities. One can write about this for hours. The BMAC folk had several animal cults which are not substantiated in any of the known IE lexicons (if one sets aside the likelihood these were loaned directly into Indo-Iranian). Specific example being the snake.//
I think you are judging by the Archaeological artifacts and putting them as cult items though they can easily be art items with nothing to do with any ''cult''. anyway the idea that BMAC Donated non-IE Deity and cult items in IIR is ridiculous every language branch of IE Specially the Large and Rooted like of IIR had many regional differences though common mythology persisted btw it is important to know that the source of those mythologies were of different times and of Different geography for example Avesta can be puted IMO in 1300 BC with the Yaz Culture can be associated with it and Rigveda can be assigned to the period of 2000-1500 BC and can be associated with Late Harappan Phase though some suktas are more old the cradle of Rigvedic phase ''was'' the Sothi-Siswal phase of Sindhu Civilization.
But is Rigveda the only representative of Indo-Aryan mythologies? definately not! many other philosophies existed for example the cult of Shakti and Shiva which is identified to be present in SSC items discovered though people try to identify them as non-IE which again pure wishful thinking as all the sources are Indo-Aryan for their identifications same with Jainism but only because
There are some attempts to make like Indra a non-IE Name etc which i think is pure nonsense.

DMXX said...

Davidski,

That seems especially likely to me. Historical accounts indicate the Cimmerians were the last remnant in the PIE homeland before the Scythians drove them southwards. We know they made it through to the South Caucasus (Scythians in pursuit). That, combined with centuries of small-scale continuous interaction with numerous early IE groups (proto-Anatolians, proto-Indo-Iranians etc.) fits the bill nicely. I just wonder if there's any evidence of "odd" archaic basal IE terms scattered across these North Caucasian languages to corroborate those forgotten interactions?

Nirjhar my friend,

That is evidently the case. The different IE branches, along their merry way, developed their own linguistic, material or cultural innovations. However, they also undoubtedly were affected by local cultures they came across. This is the crux of the matter with respect to the BMAC and evidently the source of our disagreement on this specific matter (among other things); were these innovations, or simply non-IE motifs picked up?

Also note I'm not the one describing them as "cults" (reporting the language used by archaeologists). You would think they would have their own astute reasons for labelling them as such.

General point here - A prevailing problem I have with counter views to the mainstream theory is the general inability or weakness to satisfactorily counter the most compelling linguistic evidence (recall the Indo-European enigma is a linguistic problem chiefly). I am yet to read a satisfactory counter argument regarding:

1) How else could PIE and Proto-Uralic share especially deep cognates with another
2) Why pII specifically has numerous loans with Uralic (all the way from pII to early Iranian dialects)
3) How the earliest layers of Iranic vocabulary in proto-Tocharian coincide with Andronovo
4) How pII shares satemisation with Balto-Slavic if their area of origins were not close to one another

These points are consistently left unanswered by those challenging the mainstream theory. The rebuttals I tend to read are a multidisciplinary word salad of things that rely on special pleading or are wholly circumstantial to the discussion at hand.

(I still have to read your link by the way)

Mike Thomas said...

@DMXX

* yes please do send me that article. My mind is an open sponge (i hope), and am happy for evidence to convince me either way.....

@ DMXX / Alberto

* The thing about the genetic evidence is: yes - there is autosomal evidence from "European" like input into south & central Asian populations. But this is minor compared to the input of central/west Asian into Europeans.

So if anything, the evidence supports a reverse scenario. BUT, the autosomal components we are picking up are pseudo-categories. Not stating they aren't real, but it is difficult to untangle what they actually mean cultural-historically.

Is this west -central asian component in Europeans some deep, Palaeolithic marker from the time of first AMH dispersal ? The answer IMO is a resolute no. That is already obvious from the data, incl my ancient mtDNA PCA. The Yamnaya group shifted massively from its EE HG ancestors to its new position.

I cannot account for it fully yet. The favoured opinion is wife-swapping with southern farmers. I don't think this is a convincing scenario. The amount of 'swinging' involved is mind-boggling: quite simply, the Yamnaya nomads would have to have only married non-local, southern wives. It must remain a possibility, therefore, that R1b-M269is actually central-west Asian.

Mike Thomas said...

@ DMXX

To further answer your questions

1) How else could PIE and Proto-Uralic share especially deep cognates with another

- As i explained to the 'expert linguists' previously on this log, The dep cognates between PIE and Uralic do not necessarily insist a steppe origin over a south Caspian one, for example. Its a mere 200km diference.

2) Why pII specifically has numerous loans with Uralic (all the way from pII to early Iranian dialects)

- as above. Its also about siciolinguistcis. There are no hard & fast rules. Its about langaueg dominance, 'openness', and chance also. Don;t forget also that there a linguistcs which propose strong PIE- Urartian, PIE - Caucasian and even some PIE- Semitic infleunces. These are oft overlooked by western scholars maybe because such scholars do not have as much expertise in these "oriental' languages.

3) How the earliest layers of Iranic vocabulary in proto-Tocharian coincide with Andronovo

- sorry bud, but that sounds like speculation. Tocharian is not attested until 6th c AD ! Any honest linguist will tell you that language can only be dated relatively, not absolutely.

4) How pII shares satemisation with Balto-Slavic if their area of origins were not close to one another

- first of all, Satemization might not have been one sweeping change, but a gradual, punctuated one. Moreover, the Satemization of Slavic and Baltic could have been affected by a late arrival of INdo-Iranian to the Balck Sea / EE, by the "Cimmerians" you refer to above, who are naught but 'pseudo-Cimmerians". The real Cimmerians didnt exist in EE, but lived just north of the Medes.

Nirjhar007 said...

Dear DMXX,
On Uralic and Indo-Iranian Talageri has some general Suggestions-
https://www.youtube.com/watch?v=WGXSKOXuDoU
Though let me tell i don't support OOI Model.
About PIE and Uralic well they also exist with PIE and others like Semetic, Sumerian etc and moreover they are not One way like of Uralic! there will be a article published on the issue quite soon I WILL LINK IT TO YOU AT ONCE ITS ONLINE:).
Satemization is not loci pointer as for Example Tokharian is in the Asian Satem area and Bangani a kentum dialect is in Indian area...
Moreover the most accepted theory is that PIE was not kentum, rather it had a particular palatal sound which evolved in kentum and śatam>satem, that's why in Sanskrit we have the palatal sibilant. Satemization still happened in Europe from Latin kentum: French, Spanish and Portuguese have become satem languages. So, apparently it's not possible to derive kentum from Sanskrit śatam, but we cannot exclude that before the evolution of śatam, we had a form more similar to Latin kentum in South Central Asia, like Bangani 'koto'. On the other hand, in a Kentum language like Greek we have also Satem forms like 'syn' for Latin 'cum' (Sanskrit 'sam').
I also have an interesting paper also for you on PIE location which is new thinking though its old:)
http://www.scribd.com/doc/150388032/The-Epicentre-of-the-Indo-European-Linguistic-Spread-Useful
Cheers...

Aram Palyan said...

There was a movemement from Iran to North of Central Asia in Neolithic times and maybe later also. It was the G1 haplogroup.

http://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0122968

Nirjhar007 said...

Thanks for bringing that Aram,
Yes I know that article and i have discussed it with some Professors who will be doing aDNA researches all over Asia!.

postneo said...

@dmxx

"To make matters worse for opposing viewpoints on this topic, wheel/wagon terminology was shown by Anthony to become more extensive with each IE group that left the putative steppe homeland (http://kinshipstudies.org/wp-content/uploads/2015/02/KinshipStudies-IndoEuropeanWheelMap-copy-2.jpg)"

The map has a problem. Its a mish mash of assumptions with modern languages and ancient ones. If you want to apply the same bench mark to all regions then the steppes should be uralic speaking or turkic speaking but not iranian.

I think horse eating was borrowed by some PIE speakers from northern neighbors.

Moreover the map indicates that the dialectal continuum is centered around Iran. I find it hard to believe that BMAC acquired wheels from the steppe and passed it to southern neighbours. since IVC and sumer already had wheels. Perhaps you mean spoked wheels

Gill said...

Chad, that's really interesting. What other components are in that run?

Gill said...

David, such a migration to the north Caucasus could explain the strong presence of Y-DNA haplogroup L1c-M357 in Chechens. North Caucasians are also the most Steppe/Yamnaya-like of any of the modern day Caucasus populations.

According to this:

http://yfull.com/tree/L-L1307/

Chechen/Arab L1c-M357 branched off from South Asia 3300ybp. I would bet the Arab L1c is descended from Caucasus populations in a post-Islamic timeframe.

L1c (even other clades of L1) is a major haplogroup in the Pakistan/Afghanistan/Northwest-India area.

It would be helpful if some European L individuals got ngs testing at some point to see if they share a similar timeline.

Davidski said...

Here's the spreadsheet from Chad's K=10.

https://docs.google.com/spreadsheets/d/1TjtY8caOyq_2Gg3Sv1N5PP7WGoWeJtZp0V68BRm1-EI/edit?usp=sharing

Helgenes50 said...

What is it ? this K=10
a new calculator developed by Chad ?

Chad Rohlfsen said...

It's just a spreadsheet. I think the EEF component is actually middle Neolithic, as the more Near Eastern farmers get Bedouin. Loschbour may get this component due to his involvement in the formation of it. Stuttgart is probably scoring in 3, as she is probably closer to first farmers and involved in all 3.

Chad Rohlfsen said...

I'm in the process of doing a supervised run with a Yamnaya and separate MA1 component. Hopefully, it can tease out real Yamnaya ancestry.

Gill said...

It looks like the North Caucasus component there is something like 50/50 or 60/40 ANE:ENF.

The SC-Asian one might be similar.

Vadim (MDLP) made a calculator which used South Asian ANE as a proxy for ANI and it peaked at around 60 in India going up to 70 in Pakistan. The SC-Asian here reminds me of it.

Interestingly "ASE/ASI" is getting split into East Asian and Oceanian which makes me very interested in this. I was just asking David about calculators which do that. Chad, would you consider releasing calculator files for us to tinker with?

Chad Rohlfsen said...

Oh, sorry. It should give a clearer picture of Yamnaya like ancestry, with the Caucasus and Euro.

Chad Rohlfsen said...

I don't know how to make calculators. I can look into it though. Let me see how the supervised runs turn out. Maybe, David could make a new K test.

Grey said...

Davidski

"But I'm still confused how the North Caucasus fits into all of this?"

A farmer advance onto the steppe, a tussle and then a *retreat* by the farmers to the nearest mountains (where the horse dudes couldn't hassle them any more) might explain it.

Do Georgians have any myths about arriving to their current location by boat over the Black Sea?

Kristiina said...

I would also assume that U5a1 has its origin in Eastern Europe, but I would not say that U5a1 has had much of an impact in India. The frequency of U5 in India seems to be 0.4%. Moreover, it seems to be even more frequent in Dravidian groups (including Dravidian tribals) than in Indo-European groups. On the basis of this, a part of U5a1 may have arrived already with the Dravidians considering that it has also spread to Bangladesh. U4 has a little bit higher frequency of 0.6%. By contrast, the frequency of U7 which is proposed as an Indo-Aryan haplogroup has a frequency of 2% and U2 which must be old in India has a frequency of 6.4%.

Krefter said...

Nice work Chad!!!

It makes sense.

Krefter said...

Upper Palaeolithic mtDNA R. I think this is coming from the Alps and is 40,000 or so years old. Maybe an off-shoot of Eurasians that was erased on K-14, MA-1, and WHG-types.

http://www.anthrogenica.com/showthread.php?97-Genetic-Genealogy-and-Ancient-DNA-in-the-News&p=80916&viewfull=1#post80916

Krefter said...

Here's more clear info. on that Upper Palaeolithic mtDNA. It's Proto-Aurignacian and apparently some people thought it would turn out Neanderthal not human.

http://www.anthrogenica.com/showthread.php?4384-Proto-Aurignacian-mtDNA-is-Homo-sapiens-not-Neaderthal

Mike Thomas said...

well done Chad

The Caucasus component is so dominant in yamnaya and CWC, and persists into BA (Unetice) and modern europeans.

How would you exactly define it geographically ?

Mike Thomas said...

Krefter

Thanks for that link.

Any news on the Bronze Age Pole ?

Chad Rohlfsen said...

Nirjhar,

I did post those numbers correctly.

Mike,

It seems like a North Caucasus / Kuban marker. Maybe, I'm mistaken though.

Chad Rohlfsen said...

Once the supervised is done, we should also see EEF lower in Europe, and raise in West Asia, as this EEF component looks much more like Mid-Neolithic.

Mike Thomas said...

Yes makes sense

If EEF is a mid Neolithic sign, then how would early neolithic be different ?

Chad Rohlfsen said...

It'll be lower than this component in Euros, because it lacks excess WHG, but higher in West Asians, since it has less Euro HG. I'm testing some things out right now. It may take a day or more to figure out. These runs take 3-8 hours, depending on the number of components. I'm going supervised on the next few.

Mike Thomas said...

Davidski

Maybe I missed it, but can you do some K15s on Kostenki and the Unetician sample from Haak ?

Nirjhar007 said...

Chad,
Sorry I got mistaken with Samara HG scores:P btw will it be possible to put Central Asian in the K10 to see how much it eats off the N Caucasus?.

Chad Rohlfsen said...

Central Asians are too recently and heavily mixed to be very informative. I have an MA1 and EHG component in a K11. I'm waiting on it to finish.

Nirjhar007 said...

N Caucasus aren't? anyway still out of experiment....

Aram Palyan said...

Grey

///Do Georgians have any myths about arriving to their current location by boat over the Black Sea? ///

I don't think Georgians had any myth about coming on boats. I think what You see is the result of a recent massive admixture event at 11th century AD. You can look at Hellenthal et al. map and see some Adyghean like (30 %) invasion in Middle ages. The real historic event could be the Abkhazians which happened in 10-11 centuries AD.

Georgian
Analysis: FullAnalysis
Number of individuals: 20
Conclusion: One date
Estimated date (95% CI): 1082CE (914CE - 1278CE)
Estimated proportion: 0.31
View: mixing coefficients

http://admixturemap.paintmychromosomes.com/

Mike Thomas said...

Yes, if possible, splitting the former 'Teal' into Caucasus and Central Asian propper will be informative.

Davidski said...

Aram,

The Hellenthal et al. study and the admixture atlas are horribly inaccurate and uninformative.

That's really the only way to describe them.

Mike,

I can't run any tests this weekend, but those samples you're wondering about are at GEDmatch. Maybe Krefter knows their IDs?

Aram Palyan said...

Davidski

Maybe. But something important happened in Georgia at this period especially in Western part because at this period there was a language shift from Greek to Georgian. There is a controversy who where those Abkhazian tribes but their invasion from North West is a historic fact. Georgians think that there were west Georgian tribes among Abkhazians. Abkhazians think that Abkhazian tribes were dominant. But whoever there were this don't change the genetic input story.

https://en.wikipedia.org/wiki/Kingdom_of_Abkhazia#cite_note-3

Mike Thomas said...

Dave
Whenever you can, it's not urgent .

Chad Rohlfsen said...

I'll see what I can do, but I have a feeling that the Caucasus and Sc Asia components meeting there is the story.

Chad Rohlfsen said...

Relative to Yamnaya, Tajiks would be related to them by a max of 30%, of their ancestry. It makes sense. Again, I'll keep plugging away.

Chad Rohlfsen said...

This Caucasus component may be better named as Ural/Uralic. It's huge in Siberians, where it makes up about 80% of their West Eurasian. Look at Tubalars and some Evens. Sorry, that I missed that.

Marnie said...

@Davidski

Why are you padding you Admixture runs with fake 100% WHG samples?

Chad Rohlfsen said...

What are you talking about, Marnie?

Chad Rohlfsen said...

This was unsupervised. You know what that means, right?

Krefter said...

@Mike,

https://www.dropbox.com/sh/cnnmb7cym9aer09/AABRffcNsOxHdCeNGE98nz8na/samples.xls?dl=0

Mike Thomas said...

@ Krefter

Thanks !

@ Chad

"This Caucasus component may be better named as Ural/Uralic"

Very interesting ........

Marnie said...

@Davidski

The Motala_HG samples in your spreadsheet at K=10, the one you posted the Google Docs link for, above: Why are all the Motala_HG samples shown as 99.991% EuroHG?

Chad Rohlfsen said...

It's just a label. Hunters in Europe. So what? The program picked the components. Not us! Why constantly nit-pick, instead of offering something constructive?

Davidski said...

Marnie,

The dataset that Chad used in his run includes seven Mesolithic samples from Motala, Sweden, as well as several other Scandinavian forager samples. It seems that their homogeneity is causing the ADMIXTURE algorithm to base the so called European hunter-gatherer cluster on them, which is why they're basically scoring 100% membership in this cluster, even though we know from formal mixture tests that they are a mixture of WHG, EHG (or ANE) and probably also ENA.

Also, I think another problem is that the dataset doesn't include enough WHG and EHG samples for the algorithm to create separate WHG and EHG clusters, so it creates an SHG cluster, because the SHG genomes, as per above, are essentially a mixture of WHG and EHG.

I've discussed this issue in an earlier blog post which I'm sure you've seen.

http://eurogenes.blogspot.com.au/2015/04/the-teal-people-did-they-actually-exist.html

So I'm not sure if it's possible to get around this problem at the moment, because we simply don't have enough WHG and EHG samples. But I've suggested to Chad that he should remove all of the Motala samples, except for Motala I0017, which has the most markers, to see what happens.

I hope that helps.

Nirjhar007 said...

Mike,
// splitting the former 'Teal' into Caucasus and Central Asian propper will be informative//
Exactly.
Chad,
//I'll see what I can do, but I have a feeling that the Caucasus and Sc Asia components meeting there is the story.//
Ok but will be very interesting....

postneo said...

@davidski
"The only explanation is gene flow from Europe to India with Late Neolithic/Early Bronze Age groups via the steppe."

This does not work for u5 because it's more common in south India than the north.

Davidski said...

Everything works just fine. We're talking about a very minor part of the Indian gene pool, both in the north and south.

Chad Rohlfsen said...

The unsupervised K11 wasn't that informative, when it comes to Yamnaya and SC Asia. They just shared a Central Asian, which also peaked as the West Eurasian in Siberians. I'm trying a K12. That should be done in the next 4-6 hours.

Chad Rohlfsen said...

I've removed excess SHG samples. The thing about the K11, is that it's matching Dstats. K01 and Loschbour were equally distant to SHG, but Loschbour is much closer to LaBrana. It matches that perfectly. EEF also peaks the Near East component, with Bedouins scattered in there. It looks great for those areas. Splitting up Central Asian is my goal now.

Grey said...

Aram Palyan

"I don't think Georgians had any myth about coming on boats. I think what You see is the result of a recent massive admixture event at 11th century AD"

Yeah that was a bit of a reach.

This is just thinking out loud on my part...

a model proposing a Kartvel speaking farmer group on the steppe who retreated back into the Caucasus combined with the current geographical position of the Georgians with non-Kartvel languages further north doesn't seem plausible unless

a) they retreated *around* the pre-existing Caucasus dwellers
or
b) the northern Caucasus was once Kartvel-speaking as well but it changed in the interim through later migrations/retreats.

So the next question might be do any of the north Caucasus languages have a Kartvel substrate?

Matt said...

Re: that ADMIXTURE run, ran WE samples through PCA to see what it looked like in first 2 dimensions (74% variation)

http://i.imgur.com/2267AH2.png
http://i.imgur.com/JlziaFx.png
http://i.imgur.com/up9T6AZ.png

Similar dual clines structure to WE PCA for the most part, positions at the edges are different - e.g. EEF component would expect to rotate down towards Near East, HGs, etc. Wouldn't expect it to match perfectly as some clusters will be closer to others and that would "deform" the plot. No idea if it's getting at any real populations who existed.

Chad Rohlfsen said...

Matt,

The K11 should look a lot better. The K12 is almost complete. David should be up here in the next 5-7 hours. He can post this stuff up then. The K11 takes care of the Near East / EEF problem, as it's all one component, that peaks in EEF, with additional South Asian, Central Asian, and African in Bedouins. It really makes sense.

Helgenes50 said...

@ Chad

in K11, Loschbour, did he lose his 20 % of EEF ?

Chad Rohlfsen said...

Yes. He's just 1.6% Near East, 1.6% Oceania, and 1% South Asian. The rest is Euro HG, which is topped by K01. LaBrana comes out as 91.5% Euro, 6% South Asian, 1.5% Oceania, and .5% JuHoan. Motala HG is second in Euro HG, with some getting a pinch of Amerindian. NHG samples get 1.5-11% Near East, along with some South Asian. EHG's are 10-11% Amerindian, 16-20% Central Asian, 2-5% Beringia, and 67% EuroHG.

The K12 is almost done.

Helgenes50 said...

@ Chad

All these results make sense.
By example, La Brana, in the different calculators, always has a bit of oceanian or East asian admixtures or related. It 's maybe due to his C Haplogroup.

Thanks for your work

Gill said...

Ideally it'd need a "Teal" Gedrosian-like component and a Yamnaya-like component that suitably attracted WHG/EHG/Uralic and whatever ANE/EEF/ENF they bring with them. Then a separate Euro-centric WHG-heavy component and a separate EEF/ENF component (perhaps EEF and ENF). For South Asia, using a purely Southeast Asian or South Eurasian (ASE) and Oceanian would work better than an ANE-heavy South Asian component.

Chad Rohlfsen said...

Gill,
These runs are unsupervised. The program makes these clusters. I will do supervised runs, either after the K12 or K13, depending on looks.

Krefter said...

@Helegenes,
"By example, La Brana, in the different calculators, always has a bit of oceanian or East asian admixtures or related. It 's maybe due to his C Haplogroup."

C existed in early humans all over Eurasia, not just East Asia. C1 had been in Europe with WHG-type people for over 30,000 years before La Brana-1 was born, because K-14 had it. If C1 represents something east Asian it should be in K-14, Loschbour, Neolithic Hungarians, and K01.

C1 is almost extinct in Europe today while the WHG-component is rich because of Neolithic farmers, and then Late Neolithic/Bronze age people(especially R1 rich) who erased most of the old paternal lineages.

Matt said...

@ Chad, btw - I used a formula to "split" 0.6 of the EEF into NearEast and 0.4 into Euro_HG and it generated this PCA -

http://i.imgur.com/Bo1dR7y.png

Krefter said...

@Davidski,

Which early farmers look like they have the most Basal Eurasian? Stuttgart?

I don't think they are 100% from Neolithic West Asians or even Balkans. I think this because of their high amount of Y DNA I and C-V20. They even had some U5, U4, and a U2*.

Marnie said...

@Davidski, @Chad

Could you show your admixture results at K=10, and K=11 for Stora Förvar 11 and Motala 12?

Thanks.

Chad Rohlfsen said...

Marnie,

David would probably have to run this through his K8 to see what this Euro component is. I don't know, because the software made this component.

I will try a supervised run later, to see how much EHG is in Motala, over what Loschbour may have.

Marnie said...

@Chad

Sure Chad. The software did it . . . and, no doubt, the dog ate it too.

Chad Rohlfsen said...

It's an unsupervised run, meaning it isolates the number of components I ask it to. I don't decide which samples make the components. I'll do a supervised run over the night, to show you the difference.

Davidski said...

Marnie,

I think the problem here and elsewhere is your lack of understanding of the basics as far as population genetics is concerned. As a result you become frustrated, angry and abusive.

This is not a personal insult, I'm just being honest and expressing what most of us here think.

The best solution for everyone, especially yourself, is to do a bit of research about ADMIXTURE, ADMIXTOOLS (formal statistics in general), PCA and the other methods being discussed here before commenting.

I think that once you have a better grasp of what's going on, you'll see that there's no conspiracy to fudge results, and you might then be able to offer some constructive criticism based on your knowledge of European prehistory and archeology.

Please think carefully about what I just said before responding.

Helgenes50 said...

@ Krefter

You are right. His haplogroup is probably not the solution.
That's maybe due to a low coverage of his genome

Marnie said...

@Davidski

Unfortunately for you and your collaborators, no one really cares that you are faking your data. I noticed at the SAA that Krause's paper was relegated to a minor session attended by less than 30 people. You're only hurting your field by doctoring your data ( and also libeling people online and on various genomics forums when they point out what you are up to).

Other professionals working in the field of population genetics and ancient DNA also have voiced their concern about various problems with some of these ancient DNA studies, including Tom Gilbert and Lior Pachter. Are they all "frustrated, angry and abusive" ?

Of course you are going to go on the attack to hide what you are doing. "Frustrated, angry and abusive"? Actually, I'm amused and quite busy with my own work and research to really care less about the antics of various pop gen "researchers."

Admixture is an easy tool to use. Anyone with an undergrad background in statistics could rerun your datasets, and I'm sure someone soon will. You won't be able to hide forever.

It's very apparent that you, as a man, think you can talk down to me.

There are many people who have EE/CS backgrounds who excel in population genetics specifically because the underlying algorithms are similar to those used in communication theory, stochastics and electro-magnetics. Many people coming from this background are often very well prepared, mathematically, to understand statistical packages like Admixture or Treemix.

Given my background, the fact that you are saying I don't understand how Admixture works makes it clear you are willing to demean anyone, especially a woman, who catches you in the act.

It also shed's some light on why you and your cronies never let the women, who do most of the laborious lab work on your projects, have first author on your papers. Their names are always buried in the middle.

Regardless, rest assured Davidski, that at least some other people know what you are doing. Sooner or later, it's going to catch up with you.

Davidski said...

Thanks Marnie. Glad to see we understand each other a little better now. lol

Mike Thomas said...

Weren't the latest three big papers in aDNA headed by women ? (Hungary : Nagy & Gamba; Greece: Unterlander)?

Anyhow; somone kindly previously posted a link to the 'samples' page of Pinhasi's project (I think in the "Ice Age" post). It seems to have been deleted in some of the to-ing & fri-ing. Anyone still have that link ?

Chad Rohlfsen said...

https://sites.google.com/site/pinhasierc/home

Mike Thomas said...

Thanks

Davidski said...

Apparently this paper just got accepted at Nature so we should see it within a couple of weeks. Hopefully it's not a disappointment in any major way.

Batini, Chiara et al. (2015). Large-scale recent expansion of European patrilineages shown by population resequencing. Nature Communications

Balaji said...

Thanks Krefter, Mike Thomas, Nirjhar, Capra Internetensis, Davidski and Tobus for your responses to my comments of a few days back. Regarding Davidski's idea that various U5 matrilineages found in India orginated in Eastern Europe, Kristiina has already commented that these are also found in Southern tribal and Dravidian language speakers. This makes it unlikely that they originated in Europe.

Nirjhar asked about who is considered “upper caste” in South India. Only Brahmins (about 5% of the population) are considered “upper caste”. Most people belong to the “middle castes”. I think Nirjhar is suggesting that U7 in India came from Iran in the Neolithic or Chalolithic. However the Metspalu paper is arguing against this because they have found that the U7 sub-lineages in Iran and India are different and equally ancient, dating to before the LGM.

The D statistics provided by Tobus actually do suggest a statistically significant ENA element in MA1 and Samara_HG though not in Karelia_HG and Motala_HG. D(Loschbour, MA1; Yoruba, Papuan)=0.0122 (z=2.207). This could have occurred by chance with a probability of 0.01366. D(LaBrana1, MA1; Yoruba, Papuan)=0.0079 (z=1.399), and p=0.08091. D(HungaryGamba_HG, MA1; Yoruba, Papuan)=0.0098 (z=1.686) and p=0.0459. These are independent pieces of data from three individuals separated by space and time. Therefore the probability that all three together are by chance is 0.01366*0.08091*0.0459 or 0.0000507. This is equivalent to a z of 3.887. With similar calculations, the probability that Samara_HG had the values calculated by chance is 0.000327. This is equivalent to a z of 3.408.

Lazaridis did compare the affinities of MA1 and Loschbour to different ENA populations using f4 statistics. They did not find significant differences. It appears that the D statistics are more powerful and the use of 3 different WHG individuals by Tobus has also helped.

Davidski said...

The fact that these young Eastern European-specific U5 subclades are now more common in India among Dravidian speakers doesn't mean they didn't come to India from Europe with the Indo-Aryans.

Nirjhar007 said...

David, Can you give the link of the article if there?
Balaji, Yes that's also probable.

Krefter said...

Are all African closely related like Eurasians are? Looking at Y DNA and mtDNA, Africans should all be a mix of very differnt people.

Has anyone tested variation in Africa with formal stats?

In Admixture tests Archaic humans score in African components, which makes me think African components are simply anti-Eurasian, and nothing very specific.

Mike Thomas said...

Good point

Chad Rohlfsen said...

Krefter,

I think at a K20, there is 4 separate African clusters, Mbuti, Yoruba, Ju_hoan_North, and something else which escapes me. So far, up to K12, I only have two actual SSA clusters.

Chad Rohlfsen said...

The K12 is completed. The hunter stuff looks a little better.

KO1 still leads the Euro HG
99.8% Euro HG
0.2% Oceania

Loschbour
97% Euro HG
1.9% Oceania
1.1% South Asian

LaBrana1
91.4% Euro HG
6% South Asian
2.2% Oceania
0.4% Ju Hoan North

MotalaHG
94.4% EuroHG
2.7% South Amerind
1.6% Oceania
1.3% Beringia

Ajv58
88.5% EuroHG
4.2% Central Asian
3.1% Meso Amerind
l.9% South Asian
1.5% Oceania
0.8% Ju Hoan North

KareliaHG
62.2% EuroHG
18.9% Central Asian
10.6% Meso Amerind
5.6% Beringia
2.7% South Amerind

SamaraHG
62% EuroHG
22.8% Central Asian
12.1% Meso Amerind
2.7% Beringia
0.4% South Amerind

Krefter said...

Can someone help me understand D-stats?

https://docs.google.com/document/d/1IsrAHYMhDjQ11gZwg9oI44dEkcLeIqwpgO_aT3CddF8/edit

Does a positive D-stat mean W is more related to Y than to Z and X is more related to Z than to Y.

And does a negative D-stat mean W is more related to Z than to Y and Z is more related to Y than to Z?

And if the Z-score is negative and the D-stat is negative does that mean the D-stat score is solid? Ad if the Z-score is positive and the D-stat is positive does that mean the D-stat score is solid?

Krefter said...

Chad,

Are you trying to get all Euro_HGs to score 100% in the same component? They're break into three subgroups so of course this won't work.

Chad Rohlfsen said...

Usually, a Z score under 3 is not significant. Some people think under 4 is not significant.

If it's negative, W and Z or X and Y are closer.

If it's positive X and Z or W and Y are closer.

Chad Rohlfsen said...

I'm not trying to do anything. These are unsupervised runs. The program picks and makes the number of components that I pick.

I haven't done any supervised runs yet.

Chad Rohlfsen said...

The K12 added a Meso Amerind component, on top of the Amerindian that was headed by the Piapoco and Karitiana. I renamed that one as South Amerind. I'm going up in K values to see if the Central Asia breaks into a heavily Yamnaya one, again.

Marnie said...

@Chad

"The K12 added a Meso Amerind component, on top of the Amerindian that was headed by the Piapoco and Karitiana"

Do you think that perhaps the appearance of these Amerindian components at K=12 (Admixture) might be artifacts due to the relative recent isolation of these populations?

The Anzick paper also showed the appearance of these components at higher K values:

http://linearpopulationmodel.blogspot.com/2015/01/ancient-north-eurasians-north-south.html

However, I am not sure that they signify anything other than relatively recent in place isolation of some Amerind populations.

So once these localized isolate population appear, it's unlikely that increasing the K value in Admixture will be able to tell you more about ancient human population movements.

Marnie said...

@Chad

"I'm not trying to do anything. These are unsupervised runs. The program picks and makes the number of components that I pick."

Admixture works from the data that you feed it.

Chad, could you give us a link to the sources for the Motala_HG samples (I0011, I0012, I0013, I0014, I0015, I0016, and I0017)? Are these SNPs available for download?

Thanks.

Chad Rohlfsen said...

I'm only increasing the K value to split up the Central Asian, to find something to grab onto, when doing a supervised run. I may just skip that part and go with a Yamnaya and SC Asian group as separate components.

I'm not concerned about splits in Amerindian, East Asian, and African stuff. I am looking for the best groups to grab onto for a supervised run. I will probably have a WHG/Loschbour, EHG/Karelia, EEF/Stuttgart/KO2, and then something of a Siberian, South Asian, East Asian, SSA, and possibly something West Asian. I want to see how they all behave before I pick which ones max out a component.

I think the SNP's are available at the Harvard Med site.

Chad Rohlfsen said...

I've got over 2100 samples from probably about 250-300 pops. This should give pretty clear components.

Chad Rohlfsen said...

It's not like it doesn't have a lot to choose from.

Marnie said...

@Chad

"I'm only increasing the K value to split up the Central Asian, to find something to grab onto, when doing a supervised run. I may just skip that part and go with a Yamnaya and SC Asian group as separate components."

There's no reason to think that the components that appear at higher K values for some Central Asian populations are also not artifacts of isolation. They are unlikely to be signatures of ancient population movements.

" I am looking for the best groups to grab onto for a supervised run. I will probably have a WHG/Loschbour, EHG/Karelia, EEF/Stuttgart/KO2, and then something of a Siberian, South Asian, East Asian, SSA, and possibly something West Asian. I want to see how they all behave before I pick which ones max out a component."

There's nothing to "grab onto" at these higher k values except local isolation.

Virtually all populations are linear combinations of at least pre-Neolithic gene flows (or older).

For instance, "Yamnaya" on your K=10 run is a linear combination of approximately 60% "North Caucasus", 30% "EuroHG" and between 3 and 10% "Amerindian."

These components that make up "Yamnaya" are likely biproducts of gene flow movements going back tens of thousands of years.

Other populations of Eurasian, over a very wide range, have very similar Admixture component ancestry to Yamnaya. You cannot assume that Yamnaya is specifically ancestral to other Eurasian populations using this method.

Thinking that you can isolate specific Admixture components, by manipulating the datasets you feed Admixture for various populations, is fundamentally and deeply flawed in its methodology.

Chad Rohlfsen said...

The point of all this is to mainly focus on Europe. I'm not concerned as much with scores from other places. The Steppes are a source of admixture in Europe. Mixing WHG, EHG, EEF, Yamnaya, a West Asian pop, and East Asian will be very informative to the actual breakdown in Europe. I may put in a Siberian pop to eliminate guessing as to why there is excess EHG and East Asian. That is what I'm trying to do.

Instead of making assumptions, ask what my ultimate goal is.

Chad Rohlfsen said...

Just wait and see the end product, which will be a supervised test.

Marnie said...

@Chad

If you're so interested in the population prehistory of Europe, then I would appreciate a response to my request for the link to your Motala HG sources I0011, I0012, I0013, I0014, I0015, I0016, and I0017.

Are these available for download?

Are other researchers able to reproduce your results? Do they have access to the data so that they can attempt to reproduce your results?

aniasi said...

Allow me to clarify:

I understand that ANI-ASI admixture did not occur in precisely one pulse. Rather the 'event' consisted of one period of accelerated admixture between the two populations.

Nonetheless, these results undermine the last study which asserted that there were two predominant populations, and that one was primarily Kartvelian-like, while the other was extremely isolated for 60k years and most closely related to the Andamanese.

I believe that study used far too small a population sample considering the size of South Asia's population and the complexities created by caste endogamy. The presence of 'West Eurasian' female lineages with Indian coalescence ranges in the Paleolithic, if not Epipaleolithic, means that the idea of a semi-unified and isolated ASI needs to be thrown out of the window.

I think there is some data on this population, even if somewhat accidental. A few years back, the Harappa ancestry project managed to map an indigenous 'South Asian/Indian' component that emerged when both Gedrosian/Balochi and Onge/Andmanese were filtered out. It seems to peak amongst the Kalash, Gujaratis and certain middle and upper castes in South India.

I will try and find the map.

aniasi said...

Here it is:

http://www.harappadna.org/2011/05/ref3-harappa-maps/

Chad Rohlfsen said...

I said it's at the Harvard med site. I think you can just type in AdmixTools in google search to find the Harvard Med site with the samples. It's under the downloads. Pop files are there. I'm not infront of a computer right now.

Nirjhar007 said...

aniasi,
Thanks for the suggestions:) will talk later....

Marnie said...

@Chad

Thanks. I'll contact Harvard Medical School.

Chad Rohlfsen said...

I'm home now. Why contact Harvard, when you can go here? These are available to everyone.

http://genetics.med.harvard.edu/reich/Reich_Lab/Datasets.html

Krefter said...

Chad, do these two links have everything needed to do D-stats and admixture?

http://genetics.med.harvard.edu/reichlab/Reich_Lab/Software.html

http://genetics.med.harvard.edu/reich/Reich_Lab/Datasets.html

Does an ordinary laptop have enough memory or whatever?

Chad Rohlfsen said...

Any PC is fine. You'll have to install Ubuntu/Linux on it. Dstats can be found in AdmixTools. Those links have what you'll need, as far as software and datasets.

If you want to do admixture analysis, you'll need Admixture. I think it is on the UCLA site. David can give a better idea of what you'll need. I'm still learning here.

Marnie said...

@Chad

Thanks for this link:

http://genetics.med.harvard.edu/reich/Reich_Lab/Datasets.html

Regarding Ubuntu and Linux, I already run it for some apps.

Krefter said...

@Chad,

Can you create an unsupervised K=3 ADIXTURE test with 1 or more LN/BA(BBC, UNETICE, or Urnfield) genomes, 1 or more Middle Neolithic genomes, and 1 or more Cyproit genomes.

I don't want CWC as one of the LN/BA references. The only people run through the tests can be Europeans. And maybe ignore Euros with exotic ancestry like Finns.

Obviously Cyproit genomes represent recent East Mediterranean/Near eastern ancestry and Middle Neolithic genomes represent excess EEF ancestry.

Krefter said...

A good version of the K=3 described above specifically for West Euros might be with Tuscans replacing Cyproits.

Chad Rohlfsen said...

You won't have enough WHG for that to work, in my mind. It might be a couple days before I can get to that. I have many tests in mind. My K13, unsupervised is about done, then I'm moving on to other stuff.

«Oldest ‹Older   1 – 200 of 263   Newer› Newest»