Big deal of 2016: the territory of present-day Iran cannot be the Indo-European homeland

It's been a pretty big year. Not as big as many of us had hoped for, but there's still a few more weeks till 2017, so who knows what will happen?

In any case, for me 2016 will be the year in which we finally saw hard data - courtesy of Lazaridis et al. and Broushaki at al. - that crossed off the territory of present-day Iran from the list of potential Proto-Indo-European (PIE) homelands.

Now, it's true, Lazaridis et al. and Broushaki et al. were somewhat vague in what their data meant in this context, and indeed, both sets of authors left open the possibility that what is now Iran might prove to be the PIE homeland. However, their data left no doubt: look elsewhere for the PIE Urheimat.

Why? Here it is, in point form, as simply as I can put it:

- the Mesolithic and Neolithic peoples of the Zagros range and South Caspian region were highly distinct both in terms of genome-wide genetic structure and uniparental markers, and clearly only contributed meaningful gene flow to South Asia, not to Europe or Anatolia

- after the descendants of Neolithic farmers from the Zagros range, or at least their very close relatives, migrated to South Asia, the territory of present-day Iran, as per the data in Lazaridis et al. and Broushaki et al., saw waves of migrations from the west and north that dramatically shifted the population structure of the region, bringing it closer in this respect to the Levant and Europe

- most of the Neolithic samples from Iran in Lazaridis et al. and Broushaki et al. came from near the proto-Elamite homeland in the Kor River basin in central Fars, which strongly suggests that their close relatives who streamed into South Asia were not Indo-Europeans, but rather speakers of languages closely related to Elamite

- even though Lazaridis et al. successfully modeled Early/Middle Bronze Age steppe populations, including Yamnaya, as largely of Iran Chalcolithic origin, this appeared to be a coincidence, because the Chalcolithic samples from Iran showed fairly typical South Caspian uniparental markers, such as Y-chromosome haplogroups J and G1 and mitochondrial (mtDNA) haplogroup U7, which are conspicuous by their absence from an exceptionally wide range of Bronze Age steppe samples.

I do realize that many readers won't accept these arguments for emotional reasons, because the PIE homeland debate is an emotional one for a lot of people. Nevertheless, if you decide to argue with me in the comments, make sure you have a coherent argument.

Actually, the territory of present-day Iran has never been a serious contender in the PIE homeland debate. This is something that many Iranian scholars will freely admit.

Not only was central Fars the proto-Elamite homeland, but much of what is now western Iran was the stomping ground of the Hurrians. The Indo-Europeans, in the form of Indo-Iranians such as the Medes, only got there late in the game, in all likelihood from the Eurasian steppes.

But a lot of newbies to the PIE debate don't know this, or they don't want to know it. I guess to them Iran makes sense because it's in between Europe and South Asia? Or maybe it's the fact that the word Iran is kind of similar to the word Aryan? No idea, really, but like I say, it's now all over.

See also...

A homeland, but not the homeland

Zarathushtra and his steppe posse

The story of mtDNA haplogroup U7

Days of high adventure

I've redesigned and streamlined my Principal Component Analysis (PCA) plot of West Eurasia in anticipation of the arrival of many more ancient samples. Rumor has it we'll not only get stuff from the Balkans, but also finally from the steppes north of the Black Sea and South Asia.

I'd say my new plot does a better job of highlighting relationships between the different prehistoric groups and population shifts across space and time. The datasheet is available here.

It should be pretty clear from this plot how the modern-day European gene pool came about. So I don't expect any major surprises when the new samples come in. Nevertheless, the wait is killing me, and many others I'm sure.

Update 13/11/2017: A year on, I've acquired many new samples and, as a result, streamlined the PCA some more, see...

Who's your (proto) daddy Western Europeans?

Caste is in the genes

I've built up a pretty good record over the past couple of years of pre-empting results that eventually show up in high end scientific journals.

Here's another major finding that you'll see at some point in Nature or Science: in present-day India upper caste membership is strongly associated with Eneolithic/Bronze Age steppe-related ancestry, while low caste membership and Dravidian languages are strongly associated with indigenous South Asian and Zagros Neolithic-related ancestry.

As far as I can see, with my somewhat limited dataset, the best way to show this is by plotting D-stats of the form D(Outgroup;Samara_Eneolithic)(Onge;X) versus D(Outgroup;Iran_Neolithic)(Onge;X). The relevant datasheet is available here.

Like I say, my dataset is still not great for this type of analysis, but the results are very striking and I have no doubt that they will be confirmed with more comprehensive sampling of Indian populations.

If they are confirmed, this will probably mean three things: 1) the caste system was introduced into the Indian subcontinent from the Eurasian steppe after the demise of the Chalcolithic Harappan civilization, 2) Zagros Neolithic farmers and closely related Neolithic/Chalcolithic populations of South Asia were not Indo-European-speaking, 3) like the caste system, Indo-Aryan (and thus Indo-European) languages were introduced into the Indian subcontinent from the Eurasian steppe after the Chalcolithic.

See also...

On the doorstep of India

Indian confirmation bias

The Out-of-India Theory (OIT) challenge: can we hear a viable argument for once?

Global 10: A fresh look at global genetic diversity

Update 06/02/2018: I'm working on a new version of this test. It's called the Global 25. See here.

...


Below is my new Principal Component Analysis (PCA) or genetic map of global human population structure. I think it's a little bit special, and we can discuss why in the comments if anyone's interested. The datasheet is available here; it can be used to generate 2D and 3D PCA plots, and to model samples of your choice using the nMonte and 4mix R scripts. A similar sheet with average values for the ancient populations is here.

Here are a few examples of nMonte mixture models for highly drifted populations that often confuse the crap out of standard population genetics programs.

Karitiana
Ulchi 64.45
AfontovaGora3 34.2
Dai 1.35

distance%=0.5479/distance=0.005479

Wichi
Ulchi 66.8
AfontovaGora3 33.2
Dai 0

distance%=0.5055/distance=0.005055

Kalash
Iran_Neolithic:I1945 38.6
Paniya 20.2
Yamnaya_Samara:I0357 17.4
Afanasievo:RISE509 16.55
Andronovo:RISE505 3.8
Iran_Late_Neolithic:I1671 3.45
Iran_Hotu:I1293 0

distance%=0.3793/distance=0.003793

Brahmin_Uttar_Pradesh
Paniya 54
Yamnaya_Samara:I0357 24.55
Iran_Neolithic:I1945 21.45
Afanasievo:RISE509 0
Andronovo:RISE505 0
Iran_Hotu:I1293 0
Iran_Late_Neolithic:I1671 0

distance%=0.997/distance=0.00997

Brahmin_Tamil_Nadu
Paniya 57.2
Iran_Neolithic:I1945 27.7
Yamnaya_Samara:I0357 15.1
Afanasievo:RISE509 0
Andronovo:RISE505 0
Iran_Hotu:I1293 0
Iran_Late_Neolithic:I1671 0

distance%=1.0312/distance=0.010312

Pathan
Paniya 24.9
Iran_Neolithic:I1945 23.15
Andronovo:RISE505 19.95
Iran_Late_Neolithic:I1671 17.85
Yamnaya_Samara:I0357 14.15
Afanasievo:RISE509 0
Iran_Hotu:I1293 0

distance%=0.5729/distance=0.005729

It's interesting to note that the Dai from southern China help to improve the fit for Karitiana from the Amazon basin, but not the Wichi from Argentina.

Also, Andronovo significantly improves the fit for the East Iranian Pathans or Pashtuns, but clearly not as much for the Indo-Aryan Kalash, and none at all for Brahmins from India, who are also Indo-Aryans. Why? Don't know, but it might well be an important question in regards to the origins and spread of Indo-Iranian languages.

Hurrians and the others

Here's another graph based on my new D-stats datasheet. The contrast in the population affinities of Armenia_MLBA (Middle Late Bronze Age) and Armenia_EBA (Early Bronze Age) is, at least for me, surprising.

Armenia_EBA or Kura-Araxes shows strong affinity to Caucasus populations, particularly those from the Northeast Caucasus. This is very cool, and it makes a lot of sense, because historical linguists and archaeologists generally consider Kura-Araxes people to have been early speakers of Hurrian, an ancient language thought to be closely related to present-day Northeast Caucasian languages.

But what's going on with Armenia_MLBA? I really didn't expect to see Latvians and Swedes sitting near the top of this graph. Clearly, someone from the north, closely related to present-day people from around the Baltic Sea, moved into the Armenian Plateau during or just before the Middle Bronze Age. But who were they?

I don't have a clue, but f4-stats suggest that they may have also been closely related to the Sintashta people of the Middle Bronze Age Ural steppes, who do appear very Northern European in terms of genome-wide genetic structure. The time frame fits, so does the expansive and militaristic nature of the Sintashta Culture.

Yoruba Sintashta Armenia_EBA Armenia_MLBA f4 0.000676 Z 1.395
Yoruba Potapovka Armenia_EBA Armenia_MLBA f4 0.000459 Z 1.275
Yoruba Corded_Ware_CE Armenia_EBA Armenia_MLBA f4 0.000383 Z 0.933
Yoruba Poltavka Armenia_EBA Armenia_MLBA f4 0.000373 Z 0.814
Yoruba Andronovo Armenia_EBA Armenia_MLBA f4 0.000214 Z 0.486

By the way, these stats are based on transversion sites only to limit the effects of post-mortem damage on the ancient samples, some of which are not UDG treated.

Update 06/10/2016: As far as I can see, the qpAdm modeling software shows that Sintashta is indeed the best available proxy for the European-like admixture in Armenia_MLBA.

Outgroups
Andamanese_Onge
Bichon
Chukchi
Han
Israel_Natufian
Karitiana
Kostenki14
MA1
Mbuti
Papuan
Ust_Ishim

Armenia_MLBA
Armenia_EBA 0.799±0.069
Sintashta 0.201±0.069

chisq 7.181 tail prob 0.618257

Armenia_MLBA
Armenia_EBA 0.835±0.068
Andronovo (3) 0.165±0.068

chisq 9.549 tail prob 0.388179

Armenia_MLBA
Armenia_EBA 0.842±0.065
Andronovo (4) 0.158±0.065

chisq 9.742 tail prob 0.371809

Armenia_MLBA
Armenia_EBA 0.838±0.069
Srubnaya 0.162±0.069

chisq 9.993 tail prob 0.351059

Update 08/10/2016: By the way, the fact that Kura-Araxes shares high genetic drift with many Indo-European-speaking Southern Europeans, such as Albanians, has no bearing on its posited identity as an Hurrian-speaking population. That's because this inflated genetic affinity is mediated via ancient groups of largely Near Eastern origin not directly related to Kura-Araxes, such as Copper and Bronze Age pre-Indo-European Europeans. On the other hand, Caucasians, particularly Northeast Caucasians, in all likelihood do share direct ancestry with Kura-Araxes.

See also...

On Gamkrelidze-Ivanov's dubious map skills and then some

Estonian Biocentre Human Genome Diversity Panel (EGDP)

Published along with Pagani et al. 2016, the EGDP dataset is freely available at the Estonian Biocentre website as VCF and PLINK binary files here. It overlaps at ~550K SNPs with Harvard's Human Origins, and at an impressive ~1.1 million SNPs with the ~1.2 million SNP ancient DNA chip used by the Reich Lab and others.

To see what's what, I ran a Principal Component Analysis (PCA) of all of the samples except the Congo Pygmies. I then removed four Siberians that behaved as if they had very recent European ancestry, and reran the PCA. Below are a few screen grabs from the latter analysis. The datasheet is available here.

I also tested a few of the samples with my Basal-rich K7 model. You can see their results in the spreadsheet here. Look for the individual IDs with the GS prefix.

Using the K7 spreadsheet and nMonte, here's a model for Ashkenazi Jews with some of the new EGDP populations as references, including Avars from the North Caucasus and Arabs from Israel. The Arabs do help to improve the fit, but they're not as important as Samaritans and Tuscans.

Ashkenazi
Anatolia_ChL 7.9
Arab_Israel_1 15.65
Avar 0.6
Bashkir 0.05
Cossack 0
Italian_Tuscan 30.45
Polish 11.75
Samaritan 33.6
Uygur 0

distance%=0.2874 / distance=0.002874

In any case, that looks like a fairly sensible outcome, considering that it only took me a few minutes to put together. I've seen much worse in scientific literature.

Ust'-Ishim man x2

Just wanted to see what would happen if I split Ust-Ishim into two homozygous sequences, and ran him in a Principal Component Analysis (PCA) alongside other fully homozygous individuals and composites, both modern day and ancient. In theory this shouldn't be a problem, considering that his genome is of such an incredibly high quality (~42-fold coverage). The datasheets for the PCA below are available here and here.

This is the only way that I can think of to run a PCA in which genetic drift specific to the Ust'-Ishim population of ~45K YBP western Siberia has a clear impact. Indeed, lots of interesting things on these plots. Note, for instance, that on the second plot the Afanasievo and Yamnaya-Catacomb composites appear to be attracted to the Caucasus Hunter-Gatherer composite. Also note how unusual the Neolithic Iranians (Iran_EN) appear.

Update 02/09/2016: Interactive 3D versions of these PCA are now available at Open Genomes here and here. Below are screen caps of angles that appear to recapitulate the geography of Eurasia. See comments below for details.

See also...

Ancient vs modern day West Eurasian variation

Ancient vs modern day West Eurasian variation

The Principal Component Analyses (PCA) with ancient samples that I post on this blog are amongst the most accurate and best examples of their kind that you'll see anywhere. That's not just wishful thinking; it's a fact.

My PCA don't suffer from projection bias or shrinkage, which is a handicap of PCA in many ancient DNA papers, and they're run only on observed (rather than imputed) genotypes.

However, even my PCA are far from perfect, because they're based entirely on present-day variation. In other words, I still project the ancients onto eigenvectors computed with modern day reference samples. I guess that's the equivalent of putting the cart before the horse, when originally the horse may have been a donkey, or something like that.

Nevertheless, it's the only sensible way to plot heavily degraded ancient samples with a lot of missing data. But it does often leave me wondering whether the output says anything useful about the ancient world?

Thanks to the recent release of a lot of fairly high quality ancient genotype data from West Eurasia (most of it freely available at the Reich Lab website here), I can now test how well my trademark PCA of ancient West Eurasia reflects reality.

Below are two PCA featuring ancient composite samples. The first PCA is based on ~650,000 SNPs, with 100% call rates in each of the composites. For the second PCA I pruned the markers to correct for LD or linkage, and also made sure that about half of the SNPs were from transversion sites, which are less likely to be affected by postmortem damage. That left ~125,000, hopefully relatively high quality, SNPs.

Obviously, the plots are very similar, which makes me wonder whether there's any point thinning the markers when running decent quality ancient sequences? The datasheets are available for download here and here.

Now, below is a recent example of my PCA of ancient West Eurasia. It's basically almost identical to the plots above. This is very cool, and also very important, because it means that my strategy for running PCA with ancient samples produces solid and relevant results.

Interestingly, on closer inspection, the distance between the western and eastern Neolithic farmers on the first two plots appears bloated. Conversely, the distances between the northern Hunter-Gatherer (HG) samples are somewhat reduced. Any ideas why?

Update 31/08/2016: Open Genomes generated a 3D plot based on a new PCA datasheet that I posted in the comments. Click on the image below to check it out.

Update 01/09/2016: I added present-day samples to the PCA. Very happy with the outcome. The relevant datasheet is available here.

See also...

Ust'-Ishim man x2

PCA: Neolithic Central Anatolians

Note that the individuals from the earlier site of Boncuklu basically cluster with early Neolithic Europeans, while those from Tepecik-Ciftlik are shifted south and east, suggesting an influx of admixture into central Anatolia from perhaps eastern Anatolia and the Levant after the early Neolithic. This is in accordance with the findings of Kılınç et al. who published these genomes.

I also tested the same samples with the Basal-rich K7 (refer to the spreadsheet here). Their results appear to correlate very nicely with the PCA. However, I deleted Tep001 from the PCA plot because his PCA and Basal-rich K7 outcomes didn't match, suggesting that either one or the other, or both, were spurious. This isn't surprising, however, since Tep001 only has a coverage of 0.023x.

Citation...

Gülşah Merve Kılınç et al., The Demographic Development of the First Farmers in Anatolia, Current Biology, August 8, 2016, DOI: http://dx.doi.org/10.1016/j.cub.2016.07.057

Update 15/08/2016: Below are a few admixture f3-stats from an analysis involving the new Anatolian samples. Please note, the more negative the Z score, the more likely that the target is admixed. Also, I had to use transversion SNPs to make this work, so the Z scores aren't as imposing as they might have been with more markers behind them. I'm posting all of the outcomes with Z scores lower than -1, but it might be best to ignore anything above -2.

Boncuklu_EN + Levant_N > Barcin_N f3 -0.005525 Z -2.62 SNPs 48620
Boncuklu_EN + Natufian > Barcin_N f3 -0.004252 Z -1.34 SNPs 28893

Boncuklu_EN + Natufian > Tepecik-Ciftlik_N f3 -0.013262 Z -1.566 SNPs 4384

Barcin_N + Villabruna > LBK_EN f3 -0.003652 Z -2.685 SNPs 49325
Barcin_N + LaBrana1 > LBK_EN f3 -0.003382 Z -2.462 SNPs 53537
Barcin_N + Motala_HG > LBK_EN f3 -0.002539 Z -2.388 SNPs 57533
Barcin_N + Loschbour > LBK_EN f3 -0.003089 Z -2.272 SNPs 48728
Tepecik-Ciftlik_N + Villabruna > LBK_EN f3 -0.004176 Z -1.452 SNPs 34905
Barcin_N + Hungary_HG > LBK_EN f3 -0.001939 Z -1.32 SNPs 41610
Boncuklu_EN + Levant_N > LBK_EN f3 -0.003035 Z -1.221 SNPs 40815

Barcin_N + Loschbour > Iberia_EN f3 -0.002457 Z -1.171 SNPs 38141
Tepecik-Ciftlik_N + Hungary_HG > Iberia_EN f3 -0.004022 Z -1.039 SNPs 21848

Barcin_N + Villabruna > Hungary_N f3 -0.006408 Z -4.28 SNPs 44545
Barcin_N + Hungary_HG > Hungary_N f3 -0.005216 Z -3.355 SNPs 39808
Barcin_N + Bichon > Hungary_N f3 -0.002554 Z -1.667 SNPs 48500
Barcin_N + Motala_HG > Hungary_N f3 -0.001559 Z -1.298 SNPs 51052
Tepecik-Ciftlik_N + Villabruna > Hungary_N f3 -0.003929 Z -1.239 SNPs 31535
Tepecik-Ciftlik_N + Hungary_HG > Hungary_N f3 -0.003897 Z -1.192 SNPs 28472
Barcin_N + LaBrana1 > Hungary_N f3 -0.00179 Z -1.083 SNPs 46955

On the enigmatic early Neolithic farmers from Iran

There still seems to be a lot of confusion around the traps, including in the comments at this blog, about the genetic structure of the early Neolithic Iranian farmers.

They're certainly a unique and mysterious West Eurasian population, but I'd say the picture is generally pretty straightforward considering that they were dug up on the border between the Near East and Central Asia.

As per my K7 test, they're closely related to other West Eurasians, and especially Near Easterners, via an ancient component that appears to be a mixture of Basal Eurasian and something very similar to the Villabruna cluster (see post here and the last page of the accompanying comments).

Apart from that, they harbor a lot of AG3-related ancestry, albeit probably only distantly related. My guess for now is that this is mostly admixture from an as yet unsampled Central Asian forager population, perhaps with elevated affinity to Ust_Ishim (update: probably not, see here).

The graphs below are based on the datasheet available here. Like I say, these ancient Zagros farmers are unique and eastern shifted, but, at the same time, don't show the type of Southeast Asian pull that characterizes present-day South and South Central Asians.