The genetic structure of the world’s first farmers

in the doc. I am reading
Supplementary Information 9
Constraints on the origin of Ancestral North Indians
While the Early/Middle Bronze Age ‘Yamnaya’-related group (Steppe_EMBA) is a good genetic
match (together with Neolithic Iran) for ANI, the later Middle/Late Bronze Age steppe population
(Steppe_MLBA) is not.
Steppe_MLBA includes Sintashta and Andronovo populations who have been
proposed as identical to or related to ancestral Indo-Iranians as well as the Srubnaya from eastern
Europe which are related to South Asians by their possession of Y-chromosome haplogroup
R1a1a1b2-Z935.

Does the author explain what Steppe_EMBA (Early M Bronze Age) samples were used? because if they correspond to the Yamnaya culture (Haak 2015) samples , these samples all predate by far 2000 BC then they cannot correspond to the supposed invasion of the Indus Culture in 1500BC. Further more, most of them are R1b-L23 not R1a.

I suppose that Steppe_MLBA samples come from Sintasha & Andronovo Cultures (Allentoft 2015).In this case, we are little bit closer to 1500BC time frame and some of them are R1a but obviously it is not a good genetic match. Nothing is perfect.

Also the extended data fig 4 shows that all regions of india, from North to South have similar admixture of Steppe_EMBA meaning that this admixture is not relevant to Indo-Aryan populations, the Indo-European linguistic group in India.
 
Last edited:
in the doc. I am reading
Supplementary Information 9
Constraints on the origin of Ancestral North Indians


Does the author explain what Steppe_EMBA (Early M Bronze Age) samples were used? because if they correspond to the Yamnaya culture (Haak 2015) samples , these samples all predate by far 2000 BC then they cannot correspond to the supposed invasion of the Indus Culture in 1500BC. Further more, most of them are R1b-L23 not R1a.

I suppose that Steppe_MLBA samples come from Sintasha Culture (Allentoft 2015).In this case, we are little bit closer to 1500BC time frame and most of them are R1a but obviously they are not a good genetic match. Nothing is perfect.
The difference between Steppe EMBA and MLBA is that the latter contains some Anatolian/Natufian Neolithic, but EMBA only Iranian Neolithic. Check figure 1c. Perhaps MLBA tribes, coming from West with European farmer admixture, pushed out EMBA tribes out of the Steppe, some of them went south and conquered North India.

Also interesting is that they model Iberain BA with no Steppe input. Just half Anatolian/Natufian Neolithic and half WHG.
 
and why do you think so ?
Why would you think otherwise? All the recent papers Lazaridis and Lorente says so, by admixtures, numbers, PCA and graphs, and even literally:
farmers related to those of Anatolia spread
westward into Europe
Now, all samples of Anatolian Neolithic came from NW Anatolia. Let's connect the dots.
 
in the doc. I am reading
Supplementary Information 9
Constraints on the origin of Ancestral North Indians


Does the author explain what Steppe_EMBA (Early M Bronze Age) samples were used? because if they correspond to the Yamnaya culture (Haak 2015) samples , these samples all predate by far 2000 BC then they cannot correspond to the supposed invasion of the Indus Culture in 1500BC. Further more, most of them are R1b-L23 not R1a.

I suppose that Steppe_MLBA samples come from Sintasha & Andronovo Cultures (Allentoft 2015).In this case, we are little bit closer to 1500BC time frame and some of them are R1a but obviously it is not a good genetic match. Nothing is perfect.

Also the extended data fig 4 shows that all regions of india, from North to South have similar admixture of Steppe_EMBA meaning that this admixture is not relevant to Indo-Aryan populations, the Indo-European linguistic group in India.

as far as I can tell from figure 1a and 1b steppe eneolithic is Khvalynsk, steppe EMBA is Yamna/Afanasievo and steppe MLBA is Sintashta/Andronovo

this is strange, it doesn't make sense
 
Why would you think otherwise? All the recent papers Lazaridis and Lorente says so, by admixtures, numbers, PCA and graphs, and even literally:

Now, all samples of Anatolian Neolithic came from NW Anatolia. Let's connect the dots.

well, read again, it's not the same
 
well, read again, it's not the same
Wow, read again what FH said. If you are nit picking, it's never the same. Population can change genetically in time even without mixing with others, right?
 
When I said South Asian origin of Basal Eurasian I meant India or thereabouts and I think @Bicicleur posted the same thing for the same reasons.

People are mixing up dates and associating unrelated characteristic. Basal Eurasian predates farming by tens of thousands of years. To speak of the arrival of basal Eurasian farmers makes no sense.

Everything actually makes very good sense though. We see EHG start to pop up in Iran and the Zagros around the chalcolithic, which is when we started to see the Teal increase in Steppe samples. Not surprised to see R1b in bronze age Armenia as it comes with some EHG and is the perfect spot and time for the arrival of the Anatolian languages, keeping in mind that neolithic Armenia was all L2(?) Kura axes however is highly unlikely to be IE. Probably Hurrian or Caucasian or ancestral to those languages. Also as we move to the late Bronze age in Armenia we see a shift towards the steppe, which is consistent with ever increasing steppe influence as we move through the bronze age.

We'll have to see what the P* in neolithic Iran is. If it's R1b that could be a surprise to some, but it shouldn't be. R2 at this point would make more sense.

Also, as I understand it CHG/Neolithic Iranian doesn't have EHG, but indeterminate "euro"-HG. There's a difference. They say this before they throw in EHG/WHG as a proxy. This would indicate an earlier mixing with "euro" HGs, I think, which makes sense.

In rush, I could be overlooking things.
 
And actually, the admixture in ANI might make linguistic sense. Iranian actually appears to be a bastardized version of Indic that was passed down through oral tradition more so than written. So Indic being associated perhaps with earlier (less admixed?) steppe makes sense.

There's a reason the details of this particular subject are elusive. Maybe this is why. Yamnaya like came with Indic making a more direct (unadulterated) migration into India and Sintasha/Andronovo and the like represent either a later group, or the group that hung around and mixed resulting in Iranian. So Indo-Iranian would have been more Indic like that anything else.

Also, many archaeologists maintain that Yamnaya only needs to be the origin of Indo-Iranian, archaeologically speaking. Perhaps they were onto something?
 
Last edited:
as far as I can tell from figure 1a and 1b steppe eneolithic is Khvalynsk, steppe EMBA is Yamna/Afanasievo and steppe MLBA is Sintashta/Andronovo

this is strange, it doesn't make sense

How is this strange?

*EDIT* I get it, you mean it's strange that ANI doesn't appear to have MLBA Steppe. Yes this was unexpected, but I think the explanations offered work just fine.
 
As for WHG, Gravettian seems to have spread till Karain cave, NW of Antalia.
I guess there were paleo/meso I living along the Anatolian coast all the way till Antalia. From there on they could be picked up by farmers.

I have a theory about G2a2 as well, but there is no DNA proof for that, because we have no samples from there.
13 ka (1000 years after the Villabruanans) obsidian from Melos arrived in Franchtii cave along with pulse seeds and pistacho nuts from Anatolia. It would have been brought by tuna fishers in the Aegean.
12.5 ka Cyprus was discovered by HG and abondonned again soon after, to be recolonised 10.8 ka by the first farmers.
9.5 ka first fish arrived in grotta dell'Uzo in Sicily.
Appearently there were some seafaring fishermen active along southern Anatolia, the Aegean and later also southern Italy. They were also doing some horticulutre with pulse seeds.
This may be the source of G2a2.
Few of these G2a2 may have picked up cereal farming and domesticates somewhere along the Levantine coast ca 8.5 ka.
The G2a2 split up in three branches +/- 8.5 ka : NW Anatolia, Eastern Greek peninsula and Carded/Impressed ware, Corfoe.
In PCA charts all 3 populations* are close to each other, but NW Anatolia and Eastern Greek peninsula are closer to each other than to Carded/Impressed ware.

* we have no DNA from Eastern Greek peninsula Neo, but we have Sarcevo-Koros and LBK DNA

There was not that much need for ships crossing from anatolia into europe
The black sea first entered/penetrated into the sea of Marmara in 9.2 ka

The sediments and fauna record the Black Sea outflow beginning at 9.2 ka BP and the subsequent water column stratification as Marmara Sea established its two-layer circulation.
 
When I said South Asian origin of Basal Eurasian I meant India or thereabouts and I think @Bicicleur posted the same thing for the same reasons.

People are mixing up dates and associating unrelated characteristic. Basal Eurasian predates farming by tens of thousands of years. To speak of the arrival of basal Eurasian farmers makes no sense.

Everything actually makes very good sense though. We see EHG start to pop up in Iran and the Zagros around the chalcolithic, which is when we started to see the Teal increase in Steppe samples. Not surprised to see R1b in bronze age Armenia as it comes with some EHG and is the perfect spot and time for the arrival of the Anatolian languages, keeping in mind that neolithic Armenia was all L2(?) Kura axes however is highly unlikely to be IE. Probably Hurrian or Caucasian or ancestral to those languages. Also as we move to the late Bronze age in Armenia we see a shift towards the steppe, which is consistent with ever increasing steppe influence as we move through the bronze age.

We'll have to see what the P* in neolithic Iran is. If it's R1b that could be a surprise to some, but it shouldn't be. R2 at this point would make more sense.

Also, as I understand it CHG/Neolithic Iranian doesn't have EHG, but indeterminate "euro"-HG. There's a difference. They say this before they throw in EHG/WHG as a proxy. This would indicate an earlier mixing with "euro" HGs, I think, which makes sense.

In rush, I could be overlooking things.

If Kura-Araxes was not IE which is correct imo, I guess you expect that if we get more samples from Armenia EBA R1b will not be the haplogroup of the majority. And I guess you expect that R1b (this particular subclade too) will be found a little later in Central and Western Anatolia.
Is that correct?

Also, that 'indeterminate "euro"-HG', assuming that it is not EHG, can it have common origin with EHG? Could these people have had common ancestral populations? I guess yes but I don't know much. If not, why?
 
And actually, the admixture in ANI might make linguistic sense. Iranian actually appears to be a bastardized version of Indic that was passed down through oral tradition more so than written. So Indic being associated perhaps with earlier (less admixed?) steppe makes sense.

There's a reason the details of this particular subject are elusive. Maybe this is why. Yamnaya like came with Indic making a more direct (unadulterated) migration into India and Sintasha/Andronovo and the like represent either a later group, or the group that hung around and mixed resulting in Iranian. So Indo-Iranian would have been more Indic like that anything else.

Also, many archaeologists maintain that Yamnaya only needs to be the origin of Indo-Iranian, archaeologically speaking. Perhaps they were onto something?

do you mean Sintasha/Andronovo Vedic migration to India ca 3.7 ka was preceded by a Yamnaya migration to India
or do you mean the Vedic migration to India was actually a Yamnaya migration?
 
Wow, read again what FH said. If you are nit picking, it's never the same. Population can change genetically in time even without mixing with others, right?

I adressed a question to Fire Haired. The idea is to discuss and learn from it. Please don't interfere.
 
There was not that much need for ships crossing from anatolia into europe
The black sea first entered/penetrated into the sea of Marmara in 9.2 ka

The sediments and fauna record the Black Sea outflow beginning at 9.2 ka BP and the subsequent water column stratification as Marmara Sea established its two-layer circulation.

indeed sea levels were still rising
not much from these people will be found back in the archeological record
 
If Kura-Araxes was not IE which is correct imo, I guess you expect that if we get more samples from Armenia EBA R1b will not be the haplogroup of the majority. And I guess you expect that R1b (this particular subclade too) will be found a little later in Central and Western Anatolia.
Is that correct?

We already see a lack of R1b in neolithic Armenia, then it's appearance as we move into the bronze age. So yes, I would expect this to continue until we see distributions resembling what we see today. We're only missing the intermediate.

Also, that 'indeterminate "euro"-HG', assuming that it is not EHG, can it have common origin with EHG? Could these people have had common ancestral populations? I guess yes but I don't know much. If not, why?

Well it would depend on how far back you go, of course. But I do not believe we have basal eurasian in MA-1, AG3, and AG2. Someone correct me if I'm wrong.
 
do you mean Sintasha/Andronovo Vedic migration to India ca 3.7 ka was preceded by a Yamnaya migration to India
or do you mean the Vedic migration to India was actually a Yamnaya migration?

The latter, but we may be overplaying this result. Who knows what more data will tell. I believe that the only missing thing is EEF admixture.

It's late to really be Yamnaya proper, but the population would have been genetically Yamnaya in my model. And it would have began earlier than we all estimate.
 
The difference between Steppe EMBA and MLBA is that the latter contains some Anatolian/Natufian Neolithic, but EMBA only Iranian Neolithic. Check figure 1c. Perhaps MLBA tribes, coming from West with European farmer admixture, pushed out EMBA tribes out of the Steppe, some of them went south and conquered North India.

Quite funny. If so it could be thought that the later EMBA cultures were "Cordedwareized" and "R1aized"?

Well, let's add another yellow alarm.
 
I guess one person's "nitpicking" is another person's attempt at clarity and trying to get even the details right. We're all different in how we approach the material.

I came across some posts by RK at eurogenes that might clarify things. :) I hope that RK doesn't mind my lifting part of his posts. It's a compliment. :) This relates to the "ethnogenesis" of the "Anatolian Neolithic" group.

"There's some complexity with regard to 'circular admixture' between Levant_N, Natufian and Anatolia_Neolithic, and also Iran_N, CHG, and Iran_Chalcolithic, but the qpWave rank indicates that there really isn't much more complexity left once we use Levant_N and Iran_N as the bottom two corners. Anatolia_Neolithic still requires a currently unsampled population slightly further west than WHG on the WHG-EHG cline to get represented accurately, but not much else seems missing."

"Matt" made an interesting response to this:

"Less EHG ancestry than Switzerland_HG, which is modelled as having less than Loschbour. (Or conversely, less close to the WHG who are ancestral to EHG, I guess?). I don't think this will actually prove to be necessary though, and that this will all depend on what they find when they rerun and republish with the dataset from Fu et al."

I think it makes sense if we remember that the population in Anatolia was "WHG like", but not WHG.

Also,
"About Anatolia_Neolithic, I think the proportions are probably incorrect, but the source populations are probably correct. I agree that the proximate population source for EEF is Anatolia_Neolithic, but I expect some extra Iran_N would probably fit in the Danubian Neolithic on top of Anatolia_N if we tried."

He's talking about a mix of Iran Neolithic and Levant Neolithic (plus some "WHG like") admixture.

In my opinion, the specific, proximate source for at least Cardial might not yet have been sampled, i.e. it's in the Aegean or eastern mainland Greek Neolithic, and the Marmara samples are a bit different. Until we get them though I don't think there's any problem with using the ANF's as our best source so far.

As I have pointed out as well, although these early farming people were pretty homogeneous, they weren't identical.

"Previously there was a strange but very consistent tendency for all steppe groups, including yamnaya and Khvalynsk, to favour some EEF groups over others; Iceman was always the EEF pop closest to the steppe. Also some C European EEF groups were more 'Caucasus' than others, where 'Caucasus' is an ADMIXTURE component that is not very ANE-shifted despite centering there. Now that EEF is seen to be a mix of Iran_N and Levant_N, there is probably slight variation in the proportions depending on location, with Cardial farmers more Levantine and Danubian more Iran_N, I would bet."

That could be because the Danubian might derive more of its ancestry from the northwest Anatolia samples, and the Cardial ones from the Aegean group.

As to the comment about the Basal Eurasians, I don't recall mixing up dates or saying that the original Basal Eurasians were always farmers. No group was "always farmers". All groups were originally hunter gatherers. What I was speculating is that the group or groups which developed agriculture had very high Basal Eurasian percentages.

On a related note, again from RK:
"But they do get consensus figures from matching qpAdm, f4s and ADMIXTUREGRAPH, so I am inclined to trust the higher figures. We would probably need a >>80% Basal Eurasian genome to settle the figures very well, probably from the Persian Gulf or Arabia."

As to the spread of farming:
" My guess is that, as small groups of humans passed through the Neolithic 'cultural bottleneck', the deformation of both genes and memes was so strong, and the resulting society so difficult for H-G societies to replicate, that all instances of highly developed agropastoral lifestyle involved at least some vertical genetic and memetic transmission, and accompanying demographic change in the ancestry stream at some point. "

"I expect that, once the preceding middle Eastern HGs are compared against the neolithic successors, extremely strong signs of recent selection will emerge. There are sharp boundaries in the presence of small %s of Basal ancestry in reindeer herders vs Siberian HGs, which seem to suggest some pre-adaptation is necessary for sustained agropastoralism in closed cultural groups, even today."

So I and LeBrok have been saying for a couple of years.

On the migration to the north into the steppe groups:
"the proximate contributor is probably an agricultural population on the Iran_N Levant_N clinal crescent that diffused into the Caucasus and further northwards, thus tying the Basal ancestry in pre-IEs to agropastoral groups, instead of an older diffusion from Kotias HGs only."

So, it may not be Iran Chalcolithic, but it's not Kotias CHG hunter-gatherer. It's a later farming group, not a very early hunter-gatherer group.

On why Kotias is sometimes a better fit:
"Later populations favouring Kotias over Ganj Dareh in Llorente et al. is probably an artifact caused by Kotias being a mix of Iranian+EHG/WHG+Levantine ancestry, which will have the effect of pulling Kotias away from the Iran_N corner towards the centre of the cloud of populations, where most later groups lie."

This I'm not sure about, but it seems reasonable.

As to why the models don't work with M/L steppe groups like Sintashta, Andronovo, or even Srubnaya related groups he doesn't really opine much that I can find.

One reason might be that we don't yet have the genome of the proximate but related source (related to Andronovo or Sintashta, I mean). Or perhaps they have it but aren't yet ready to release it. The answers might also be clearer once we have a genome from BMAC and IVC.

"
The existence of a Basal-EHG admixed population deep into Iran also makes a very old ANE presence in Central Asia somewhat more likely, especially if the Basal ancestry peters out as one moves deeper into the North and East. My guess is that the IVC will be Iran_N+Onge and BMAC will be Iran_N+ANE/EHG, reflecting admixture from the prior HG populations in each location. "

He also dives into the controversy over Proto-Afro-Asiatic. Again, I've always thought it probably moved with Levantine people south rather than the other direction, perhaps from the Sinai with even more Natufian like people.

I do think that the "steppe" proportions in even southern South Asian tribals has to do partly with pre-existing "ANE like" ancestry.
 
As to the comment about the Basal Eurasians, I don't recall mixing up dates or saying that the original Basal Eurasians were always farmers. No group was "always farmers". All groups were originally hunter gatherers. What I was speculating is that the group or groups which developed agriculture had very high Basal Eurasian percentages.

On a related note, again from RK:
"But they do get consensus figures from matching qpAdm, f4s and ADMIXTUREGRAPH, so I am inclined to trust the higher figures. We would probably need a >>80% Basal Eurasian genome to settle the figures very well, probably from the Persian Gulf or Arabia."

As to the spread of farming:
" My guess is that, as small groups of humans passed through the Neolithic 'cultural bottleneck', the deformation of both genes and memes was so strong, and the resulting society so difficult for H-G societies to replicate, that all instances of highly developed agropastoral lifestyle involved at least some vertical genetic and memetic transmission, and accompanying demographic change in the ancestry stream at some point. "

"I expect that, once the preceding middle Eastern HGs are compared against the neolithic successors, extremely strong signs of recent selection will emerge. There are sharp boundaries in the presence of small %s of Basal ancestry in reindeer herders vs Siberian HGs, which seem to suggest some pre-adaptation is necessary for sustained agropastoralism in closed cultural groups, even today."

So I and LeBrok have been saying for a couple of years.

It makes sense that farming spreads with genes but in the Middle East there's no evidence that it did. Basal Eurasian was prevalent before farming and didn't rise in frequency once people started to farm because of natural selection.
 

This thread has been viewed 217616 times.

Back
Top