search this blog

Wednesday, May 16, 2018

New PCA featuring Botai horse tamers, Hun and Saka warriors, and many more...


Just in case anyone's wondering how the ancient samples from the two recent archaeogenetic papers by Damgaard et al. (Nature and Science) behave in my two main Principal Component Analyses (PCA), here you go:


The relevant datasheet is available here. Just over 100 of the new samples made into onto this plot, but to keep things simple I only highlighted a few of them. To see the positions of any or all of the rest, plug the datasheet into, say, PAST (freely available here) and create your own version of the plot. Also, below are links to updated Global25 datasheets, featuring coordinates for almost all of the new samples (available separately here).

Global25 datasheet ancient scaled

Global25 pop averages ancient scaled

Global25 datasheet ancient

Global25 pop averages ancient

The interesting thing about those Tian Shan nomads, especially the Kangju people, is that they're much more West Eurasian (European + West Asian) than the Asian Scythians sampled to date. However, despite this, they're still no good for modeling the West Eurasian ancestry of most South Asian populations. I've looked at this closely, and the Steppe_MLBA cluster is still the one to beat in this respect.

See also...

Genetic ancestry online store (to be updated regularly)

67 comments:

zardos said...

Do I miss something or are there no East Asians? Doesnt the lack of an EA reference distort some relations?

Davidski said...

@zardos

There's a rough threshold of East Asian admixture for the West Eurasian PCA. Outcomes for individuals that have around 30% and more East Asian ancestry aren't very informative, but below that it works fine.

The purpose of running PCA that focus on certain world regions, like West Eurasia and Northern Europe, is that they're able to flesh out more differences between the populations native to those regions.

But, in any case, the Global25 PCA analysis has samples from all over the world, including many from East Asia.

Anonymous said...

@ Salden
"It's behind a paywall and mtDNA only."
My disappointment is immeasurable and my day is ruined.

Onur Dincer said...

The interesting thing about those Tien Shan nomads, especially the Kangju people, is that they're much more West Eurasian (European + West Asian) than the Asian Scythians sampled to date.

That is understandable. Tien Shan region is geographically somewhat transitional between historically sedentary Southern Central Asia and historically pastoral nomadic Northern Central Asia even though it belongs more to the latter. Scythians/Sakas were peoples of the pastoral nomadic zone of Central Asia, so the East Eurasian admixture must have arrived much of their territories earlier than Tien Shan region.

Open Genomes said...

@David and all

Many of the ancient Eurasian samples are ambiguously grouped into populations, are outliers, or just don't belong where the studies say they should.

A good way to figure out where individuals really fit in is to do a cluster analysis a large set of samples.

Here is an analysis dendrogram based on the Ward distance-squared algorithm:

Ancient Eurasia Eurogenes Global25 clustering dendrogram

(In this PDF, zoom in all the way in and move to the right edge to see the sample names, populations, periods, and regions. The PDF text is searchable.)

Some things we can see in the tree:

• The Mycenaeans and Minoans cluster with the Bronze Age Anatolians, and the Caucasus (Armenia-Hajji Firuz) Chalcolithic

• The Sappali Tepe culture seems to have a relationship
to Hajji Firuz and Seh Gabi ChL

• The BMAC culture inlucding Tepe Hissar in Iran originates with the Namazga Neolithic, and it has a distant relationship to the Iran Early Neolithic

• The so-called "Siberia Neolithic" is actually the Botai Culture and actually groups together many of the steppe ouliers, as well as a couple of Khvalynsk Neolithic individuals

• One Early Corded Ware individual I4629 from the Baltic actually seems to be from Yamnaya

• One Comb Ceramic individual, Tamula1, clusters with the EHG Urkaine Mesolithic and Neolithic

• The XiongNu cluster with Turkic-speakers, and so does the Early Medieval Gepid from Serbia and M2198, the "Iron Age" Anatolian, who may actually be a Seljuk

• The Kangju, Wusun and Tien Shan Huns cluster with Scythians, Sakas, Sarmatians, and Alans

• One "Hungarian Scythian", DA194, actually seems to be a Bell Beaker

• Shamanka and Ust Ida Neolithic with Lokomotiv from the West Baikal are a new ancient component that may be linked to Mongolic or Tungusic speakers

I'm sure there are other interesting things you can find here.

The clustering can be useful to create actual "populations" for nMonte analysis, based on how the samples cluster rather than just the population names. This will make the nMonte analysis much more accurate, and help get rid of the overfitting problem because the "population averages" are not real averages for true populations.

Gill said...

Why did you use the Ward algorithm? Over UPGMA?

Have you tried making a neighbor joining dendogram? What similarity index do you think works best?

Unknown said...

It is interesting to see these Tien-Shan nomads are in between the contemporary populations of Volga-Ural and SC-Asia on the PCA plot.

Anonymous said...

@David

Did you ever try to model MA2203 with the Varna outlier + Anatalian_EBA?

Anacoreta said...

@Open Genomes, the dendrogram it's pretty interesting and it's worthwhile to study in detail. It is also interesting to compare it with the PCAs.

Open Genomes said...

@Gill

I used the Ward distance-squared algorithm because that's the most basic clustering algorithm.

R also has UPGMA and WPGMA available, and I can run these using the following algorithms:
"ward", "single", "complete", "average", "mcquitty", "median" or "centroid".

Roy King can elaborate more on this in a technical way, but from what I understand, the other algorithms such as UPGMA leave a lot of outliers compared to Ward distance-squared. This is very refined data, not raw genetic data (i.e SNPs) so other kinds of algorithms may just emphasize the differences.

If you think there's value in a neighbor-joining tree, I can do that as well, but first discuss the technicalities with Roy King, who is a an academic statistician.

@anb anbb

Just you wait till you see the combined ancient and modern Ward distance-squared tree! ;) There are some really remarkable things there, that raise some questions about cross-cultural contacts across long distances.

rozenblatt said...

Maykop preprint: https://www.biorxiv.org/content/early/2018/05/16/322347

Anonymous said...

Only Maykop of the Steppe (and little near late North Caucasus) haa EHG.


Maikop is not Indo-Europeans.

Final.

Open Genomes said...

@David

There are some non-unique sample IDs in the Global_25 table. (This is the part after the ":".)

Many of these appear to be the same individual who appears twice in the table.

Can you fix these to remove the duplicates or make them unique?


Azeri:azerB38
Azeri_Iran:azerB38

Azeri:azerB59,azerB59
Azeri_Iran:azerB59

Azeri:azerB61,azerB61
Azeri_Iran:azerB61

Azeri:azerB64
Azeri_Iran:azerB64

Azeri:azerb72
Azeri_Iran:azerb72

Azeri:azerB8
Azeri_Iran:azerB8

Azeri:azerE1
Azeri_Iran:azerE1

Azeri_Iran:azerE3

Azeri:azerE3
Azeri_Iran:azerE6

Azeri:azerE70
Azeri_Iran:azerE70

Ethiopian_Ari_blacksmith:ARI11,ARI11
Ethiopian_Ari_cultivator:ARI11

Ethiopian_Ari_blacksmith:ARI5
Ethiopian_Ari_cultivator:ARI5

Ethiopian_Ari_blacksmith:ARI6
Ethiopian_Ari_cultivator:ARI6

Ethiopian_Ari_blacksmith:ARI7
Ethiopian_Ari_cultivator:ARI7

Finnish:HG00171,HG00171
Finnish_East:HG00171

Greek_Central_Anatolia:G25001
Greek_Trabzon:G25001

Greek_Central_Anatolia:G25002
Greek_Trabzon:G25002

Greek_Central_Anatolia:G25003
Greek_Trabzon:G25003

Indian_K:K-126

Pakistani_Kohistani:K-126

Mordovian:495_R01C01
Mordovian:495_R01C02
Mordovian:495_R02C01
Mordovian:495_R02C02


Uzbek:495_R01C01
Uzbek:495_R01C02
Uzbek:495_R02C01
Uzbek:495_R02C02

vAsiSTha said...

@supernord
Lol

Anonymous said...



Catacomb culture is not Proto-Aryans.

Matt said...

Off topic (and not as interesting as Maykop!) but: http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0196335

"Earliest evidence for equid bit wear in the ancient Near East: The "ass" from Early Bronze Age Tell eṣ-Ṣâfi/Gath, Israel"

Suggests donkeys being ridden as early as 2800 BCE, overlapping with tail end of Yamnaya period. Even assuming that this is does not imply earlier finds that we don't know about yet, this seems to suggests to me that the idea of riding equids may well have been generally "about", and that the arguments for non-horse riding by early groups on the steppe may be on weakened footing.

Anonymous said...

@Mr. Kulkarni You are LOL that ignoring facts.

Archaelog said...

The Y chromosomes are dominated by R1b. Which subclade though?

Anonymous said...

@Matt

You know, about any smart thins adopted by HG's is developed elsewhere. But smart HG's that come int contact with farmers adapt very fast. Ertebolla culture traded pigs with LBK, as their earliest pigs remains have a Anatolian mtDNA affinity. Soon afterwards the pigs show a local mtDNA affinity. And late LBK's pigs also show a local affinity. It goes to show how fast HG cultures can adapt. The American Indians adapted to horses within a generation. Back when racist alcohol laws had the unintended by-effect to protect Aboriginals from alcoholism Aboriginals outcompeted everyone at remote cattle station for cattle men.

Pastoralism is the natural extension of HG's, but wasn't their invention.

Samuel Andrews said...

About the origin of Iranian and Indo Iranian languages....

An southern Iran neolithic heavy origin isn't consistent with the genetic data. The only thing that strongly links all Iranians is Andronovo ancestry.

Kangju and modern Tajik are basically a two-way Andronovo-BMAC mix. Siberian Scythians are basically a two way Sintashta-ancient Siberia mix.

A Siberian origin for Iranian languages doesn't make sense because so many ancient and modern Iranians lack Siberian admixture. A BMAC origin doesn't make sense because Siberian Scythians lacked (more than negligible) BMAC admixture.

Karl_K said...

R1b1a2
R1b1a2a2

Anonymous said...

Chetan said...
" The Y chromosomes are dominated by R1b. Which subclade though?"

Non dominated. In Steppe Maykop and North Caucasus culture, that it is not(!) Maykop, there are R1b1a1a2(a), from Yamnaya Caucasus culture.
Maykop is not R1b, but is
J2a1
G2a2a
L

Romulus the I2a L233+ Proto Balto-Slav, layer of Corded Ware Women said...

The steppe groups from Yamnaya and subsequent pastoralist cultures show evidence for previously undetected Anatolian farmer-related ancestry from different contact zones, while Steppe Maykop individuals harbour additional Upper Palaeolithic Siberian and Native American related ancestry.

Wowzers

Rob said...

Wow amazing paper
So it’s confirmed the source of ChG admixture and acculturation of the steppe was from Majkop interactions c 4500 BC. I’m shocked !

No magic CHG sanctuary requires

This steppe Majkop group then expanded and replaced Khvalynsk ! Sorry Supernord

Arza said...

@ Rob
No magic CHG sanctuary requires

Our fitted model recapitulates the genetic separation between the Caucasus and Steppe groups with the Eneolithic steppe individuals deriving more than 60% of ancestry from EHG and the remainder from a CHG-related basal lineage, whereas the Maykop group received about 86.4% from CHG, 9.6% Anatolian farming related ancestry, and 4% from EHG.

Dunno, but "CHG-related basal linage" in Steppe EN contrasted with plain CHG in Maykop sounds like a magic CHG sanctuary.

Matt said...

@epoch2013, well, I would not make claims about directionality necessarily, it seems more like the evidence of donkey riding strengthens the case that riding was an idea that was about, or was easily adopted. not so much that it came from a particular direction.

@Romulus, yeah, some of these Steppe Maykop individuals ("individuals from the eastern desert steppes") seem to basically have West_Siberia_HG ancestry, which Yamnaya and "Eneolithic_Steppe" don't really have (or at least any significant amount of). Though they don't actually call it that, as they lack the samples from Namasimran and Damgaard.

These individuals later got replaced, or did not contribute significantly to Afanasievo? It's not a change to our idea of the Yamnaya's ancestry, but rather showing that the West Siberia HG ancestry (and Botai ancestry, really) was about a bit more than thought.

Unknown said...

So synthesizing all the new data (linguistics, archaeology, genetics), the picture that is emerging is the following:

https://i.imgur.com/Wyta11s.jpg

Rob said...

@ Arza
Nope . Just differential streams and gradients of ANF-related ancestry and CHG/ Iran
No magic source of CHG females

Also, “This implies an overlap of symbols with a communication and interaction
626 network that formed during the late 4th millennium BCE and operated across the
627 Black Sea area involving the Caucasus59, and later also involved early Globular
Amphora groups in the Carpathians and east/central Europe61. The role of early
Yamnaya groups within this network is still unclear. However, this interaction zone
pre-dates any direct influence of Yamnaya groups in Europe or the succeeding
formation of the Corded Ware”

Yamnaya - CWC are just late phases, homogenisation and spread to the Mesolithic Baltics etc
Tail end

Anonymous said...

@Rob

"So it’s confirmed the source of ChG admixture and acculturation of the steppe was from Majkop interactions c 4500 BC. I’m shocked !

This steppe Majkop group then expanded and replaced Khvalynsk ! Sorry Supernord "


you write nonsense and complete lies completely contrary to the data and the paper!

Rob said...

@supernord
Nonsense writes you
Shaaaammmeee

Matt said...

"The genetic prehistory of the Greater Caucasus": Based on PCA and ADMIXTURE plots we observe two distinct genetic clusters: one cluster falls with previously published ancient individuals from the West Eurasian steppe (hence termed ‘Steppe’), and the second clusters with present-day southern Caucasian populations and ancient Bronze Age individuals from today’s Armenia (henceforth called ‘Caucasus’), while a few individuals take on intermediate positions between the two. The stark distinction seen in our temporal transect is also visible in the Y-chromosome haplogroup distribution, with R1/R1b1 and Q1a2 types in the Steppe and L, J, and G2 types in the Caucasus cluster (Fig. 3A, Supplementary Data 1). In contrast, the mitochondrial haplogroup distribution is more diverse and almost identical in both groups (Fig. 3B, Supplementary Data 1).

Seems to basically support Davidski's contention that admixture between northern and southern pool was sex biased, though in a slightly different formulation. Here northern pool and southern pool have same mt and different y. Rather than a founder effect giving an illusory impression of sex biased admixture. Caucasus Maykop certainly not linked to R1 of any sort.

Suggests "steppe women" also sex biased admixing into the south at a similar clip to what was happening to the north? Unless assuming no steppe female contribution to the north, in which case resolved through one way sex admix after all. Patrilocality in both south and north?

However, kind of rejects the idea that there were ever groups of Caucasus derived ancestry moving about in the steppe who were absorbed by EHG groups (which was another idea I recall being raised by David and others). No samples found in Maykop Steppe ever overlap Maykop Caucasus, though 2x outliers are somewhat intermediate.

Changeover in ancestry across the border is also quite sharp, so admixture between groups almost certainly happened across a shallow interface and would seem did not involve groups from Caucasus venturing into steppe, then later being absorbed (or defeated or other colourful terms).

@Arza, I'm certainly far from clear if they are suggesting that Maykop samples or even Eneolithic Caucasus are feasible ancestors for Eneolithic Steppe, with CHG! Must read this paper again a few more times.

Open Genomes said...

@Stephan Molyneaux

About your map:

There's no evidence from this study or any genetic evidence at all that Proto-Uralic has anything to do with Sumerian.

It's much more likely that Sumerian is somewhat similar to Late Chalcolithic Seh Gabi, because these have some Near Eastern admixture and also there's a Sumerian layer at Seh Gabi dating to about 3200 BCE. However, Seh Gabi Chalcolithic is at latest 3800 BCE, and so it's much more likely to be Proto-Elamite.

Anonymous said...

@Rob is not able to read crazy. Shame on you. Are not you ashamed?


Rob said...

@ Matt
Check suppl 6- “Dating gene flow from Caucasus mountain groups into Steppe Maykop”

Rob said...

@ Supernord
Yes I repent. Please whip me

vAsiSTha said...

@all

where is the likely origin of Y haplo L-M20 & L-M27?

Karl_K said...

My guess is that there was probably some kind of magic CHG sanctuary, of only females, along a line exactly between the Steppe and Caucasus. It fits what we know from the data.

vAsiSTha said...

With more Y haplo L being found in the caucasus region, with earlier L1A1 being found in Armenia 4000BC, i think is evidence of movement from SA to caucasus.

Matt said...

@Rob, flow from Steppe_Maykop_outlier into Steppe_Maykop looks no problem, but Yamnaya_Samara have a distinct position in this PCA and ADMIXTURE (and Yamnaya Samara from Yamnaya Caucasus). As do Samara_Eneolithic against samples labeled Eneolithic_Steppe. Not clear if Yamnaya_Samara from Samara_Eneolithic+X (European EEF? GAC_Ukraine?) without need for any contribution from Maykop groups (Caucasus/Steppe).

Samuel Andrews said...

There's not enough samples to say Maykop and Yamnaya had similar mtDNA. The U7b and U1b in Maykop are quite odd for Steppe or modern Europe.

Samuel Andrews said...

This paper uses Hungary Yamanya samples. They seem to have very little European farmer admixture. They might be the R1b L151 samples we've been looking for.

Rob said...

@ Karl

"My guess is that there was probably some kind of magic CHG sanctuary, of only females, along a line exactly between the Steppe and Caucasus. It fits what we know from the data."

So this existed before 4500 BC ? Which sites - I appear to have missed them from the series of aDNA papers to date. Crimea ? Cryptic foragers from the Volga ?

Karl_K said...

@Rob

It's just a sampling issue. They'll find it soon, I'm sure. The biggest issue is that their custom was to have an air burial followed by cremation.

Rob said...

@ Karl

"The biggest issue is that their custom was to have an air burial followed by cremation."

Ah. Just like the Hittites- the real ones, not those damned Hurrians.

Samuel Andrews said...

We all detected the AnatoliaNeo stuff in Yamnaya years ago. If PIE lacked EHG, then were they pure CHG? That seems strange. What do the supposed Hittite and Yamnaya genomes share? A slither of CHG?

Karl_K said...

Well, it's just a shame that languages absolutely require the correct DNA to understand them. But, that's just how it is.

The Hittites didn't know this, though, because they missed the memo about it.

Arza said...

@ Karl_K
My guess is that there was probably some kind of magic CHG sanctuary, of only females

2 x Y-DNA J in EHG.

Also check the model in Narasimhan et al. 2018 Fig. 2B where Steppe EMBA gets additional shot of "Iranian farmer related" ancestry over Khvalynsk_EN (and de facto over Steppe MLBA if you'll look closely). Homogenisation of mtDNA gene-pools may have occured at this stage.

Karl_K said...

@Azra

Yes! I hadn't thought of that.

Perhaps PIE was spoken by Iranian Neolithic and Indus Valley people.

Interesting...

Rob said...

@ Matt

"@Rob, flow from Steppe_Maykop_outlier into Steppe_Maykop looks no problem, but Yamnaya_Samara have a distinct position in this PCA and ADMIXTURE (and Yamnaya Samara from Yamnaya Caucasus). As do Samara_Eneolithic against samples labeled Eneolithic_Steppe. Not clear if Yamnaya_Samara from Samara_Eneolithic+X (European EEF? GAC_Ukraine?) without need for any contribution from Maykop groups (Caucasus/Steppe)."

Yes, interesting. Guess we'll have to wait for data.
So Steppe-Majkop are on outgroup to other contemporary R1b ? I think one issue is the temporal gap between Majkope 3800 -> and our Yamnaya samples 3000-2600 BC, and different social processes - the steppe during Yamnaya being a democratization of craft previously restricted to elites.
Whatever the case, the daggers, stelae, etc all appear to manifest in Majkop and expand from there, corporate groups acquiring & expanding shared norms ?

Anonymous said...


There are no traces of hypothetical movement of Proto-Hittites through Caucasus.


Karl_K said...
"Perhaps PIE was spoken by Iranian Neolithic and Indus Valley people."

Impossible.

Rob said...

@ supernord
You're missing karl's humour. It's an acquired taste.

Rob said...

What's the deal of extra Karitania affinities of steppe Majkop ? also manifest as extra Han related ancestry not explained by EHG, Khvalynsk, etc . Perhaps related to Q1a +/- L1 migrants ?

Anonymous said...

@supernord

Just saying its impossible doesn't actually make it come true.

Its a real possibility (and acknowledged as such in this paper) that PIE could have gone to IVC through a southern route. Even though it seems like they still prefer the steppe route, they have finally acknowledged the possibility of the southern route towards the IVC for the first time. My own gut feeling is that PIE comes to South Asia twice through both routes.

My guess is that when we get more data in the future, we'll find that the two groups that make up the IVC (Iran farmer and South Asian HG) came to the party with their own two separate languages (early PIE and Proto-Dravidian). And this works perfectly with the logo syllabic script at the IVC.

Now I'm sure you'll scream out outrageous or sacre bleu or maybe impossible again, but this is a legit possibility at this point.

Karl_K said...

@Rob

It's that "Siberian_HG" ancestry.

Open Genomes said...

@ Mr. Kulkarni

L-M27, which is modal in South Asia (in particular the Indus River region) seems to originate in Northern Anatolia and the Chalcolithic Fertile Crescent.

YFull L-M22 tree including L-M27

As you know, two previous Maykop individuals were in L-M27. What's interesting is that there are no L-M27s in the region today, only earlier subclades of L, which are pretty common along the Black Sea coast, Northern Syria, and Northern Iraq.

Samuel Andrews said...

@Karl_K,
"My guess is that there was probably some kind of magic CHG sanctuary, of only females, along a line exactly between the Steppe and Caucasus. It fits what we know from the data."

The Amazons!!

Unknown said...

@OpenGenomes
"About your map:

There's no evidence from this study or any genetic evidence at all that Proto-Uralic has anything to do with Sumerian.

It's much more likely that Sumerian is somewhat similar to Late Chalcolithic Seh Gabi, because these have some Near Eastern admixture and also there's a Sumerian layer at Seh Gabi dating to about 3200 BCE. However, Seh Gabi Chalcolithic is at latest 3800 BCE, and so it's much more likely to be Proto-Elamite."

Sorry, i was pointing to the linguistic proof: http://s155239215.onlinehome.us/turkic/42TurkicAndSumer/SimoParpola_Altaic-UralicAndSumerEn.htm

@Arza
"My guess is that there was probably some kind of magic CHG sanctuary, of only females

2 x Y-DNA J in EHG.

Also check the model in Narasimhan et al. 2018 Fig. 2B where Steppe EMBA gets additional shot of "Iranian farmer related" ancestry over Khvalynsk_EN (and de facto over Steppe MLBA if you'll look closely). Homogenisation of mtDNA gene-pools may have occured at this stage."

From where on this map: https://i.imgur.com/Wyta11s.jpg


Gill said...

@ Open Genomes

Would neighbor joining method be better than clustering with Ward's?

Open Genomes said...

@Gill @Roy King

According to Roy King, Ward's algorithm is better for this data because the data is already refined, not raw data. This is actually a second-order analysis, the raw SNPs going into Global 25
He says that a neighbor-joining tree will separate out outliers which doesn't need to be done here.

Roy King needs to explain this in a more technical way.

I can try it though, and see what the differences are. I'm first doing the Ward's distance-squared tree for the ancient and modern samples combined, and I can try that afterwards if it's worthwhile.

Archaelog said...

@Open Genomes Thanks for sharing the subclades

Anonymous said...

thorin23 said...
" @supernord
Just saying its impossible doesn't actually make it come true."
No, it is actually true.

" Its a real possibility (and acknowledged as such in this paper).... My guess is that when we get more data in the future, we'll find that the two groups that make up the IVC (Iran farmer and South Asian HG) "

It is really impossible. There is no word about IVC in the new paper. In others, it is written a complete refutation of your words. No one recognized the possibility that IVC is Indo-European, in any of the articles. Thus, neither the Neolithic Iran and IVC has not been recognized Indo-European. The way of Indo-Aryans in IVC through Siberia and the Mountain Corridor of Central Asia was considered. No one recognized the way through Late Bronze Iran and the Caucasus, although such a way has always been hypothetically possible without any proofs. IVC could not be Indo-Aryan, these are the conclusions of the papers. It was only about whether Indo-Aryans the destroyers of the IVC or not.

Grey said...

"“This implies an overlap of symbols with a communication and interaction
626 network that formed during the late 4th millennium BCE and operated across the
627 Black Sea area involving the Caucasus59, and later also involved early Globular
Amphora groups in the Carpathians and east/central Europe"

sounds a bit like a collection of peoples connected across the Black sea - could call the collection the "Black Sea Peoples" if most of that name wasn't already in use for something else.

Grey said...

Matt

"However, kind of rejects the idea that there were ever groups of Caucasus derived ancestry moving about in the steppe who were absorbed by EHG groups (which was another idea I recall being raised by David and others)."

yeah - farmers moving onto the steppe, catalyzing a herder culture in the region bordering the crop farming limit, and then eventually getting over-run by the herders seemed to me one of the possible ways to explain the mtdna disparity.

iirc the other was bride swapping across the farmer/herder border region followed by massive expansion of the new hybrid steppe herder population over the steppe HGs who lived away from the contact zone.

(if the density of farmer population was higher then the herder population then an equal number of brides would have a larger percentage impact on the steppe population?)

#

also wouldn't this imply the steppe dudes were like Ertobolle in adopting bits of agriculture from adjacent farmers rather than having it imposed by intrusive off-steppe farmers?

Grey said...

supernord said...

"There are no traces of hypothetical movement of Proto-Hittites through Caucasus."

with all these wetland forager HGs around the Black Sea you'd think they'd maybe have learned to fish and trade a lot - on boats

Matt said...

Grey: iirc the other was bride swapping across the farmer/herder border region followed by massive expansion of the new hybrid steppe herder population over the steppe HGs who lived away from the contact zone.

(if the density of farmer population was higher then the herder population then an equal number of brides would have a larger percentage impact on the steppe population?)


Though I sometimes find some of your populations on demographic density scenarios a bit complex (too many unknowns!), this is actually a great point Grey. *Could* sex biased admixture for a female exchange scenario even be possible against a lower->higher population density?

For males, it's clearly possible under some circumstances because of male lineage explosiveness (star like expansion qualities). It doesn't matter if the HG males are small in number at first, if they can reproduce a mechanism that favours their sons, systematically, and/or allows them to acquire multiple wives and support those children.

But for women with lower reproductive variance and shouldering the main burden and ability to reproduce our species, going sex biased admixture against the gradient of population density is gonna be really hard.

This may explain a bit of why admixture with HG males in Europe *had* to be sex biased in favour of male HGs. Male HGs just can starburst into EEF populations in a way that HG females couldn't, whether male HG:female farmer combinations were any more advantaged than male farmer:female HG or not.

Of course, you can certainly have male farmer:female HG admixture (West Africa shows an example, as, probably, does the steppe). That's not impossible at all. But unless the mix is situated to then boom demographically along a new frontier, rather than simply get absorbed back into the farmer population, it seems like it will always be demographically marginal in ultimate effect.

Grey said...

Matt
"(too many unknowns!)"

sure - not saying this is what actually happened just making an arithmetical point - in theory you could have an equal number of bride exchanges over time across a frontier between group A and group B (for example alliance marriages between the chiefs of two tribes) but *if* the population density of the two groups was different then the autosomal impact of those exchanges would vary in proportion to the relative density e.g. if tribe A had a 100 people and tribe B had a 1000 and there were 10 bride exchanges that would be 10% of A but only 1% of B - which *might* be an explanation for a larger autosomal impact on one side of a biozone border than the other.

Davidski said...

@All

I updated the above PCA plot and datasheet with new versions of the same samples. For more details see here...

Downloadable genotypes of present-day and ancient DNA data