drought of ancient DNA papers on prehistoric Europe/SW Asia

[Ryukendo]

It's ok, you can direct all your 'A Basal Eurasian and an Aurignacian walked into a bar' jokes toward me

Kale, this is getting quoted in my sig

[TuaMan]

So.... any word on any upcoming bigtime papers?

Besides Old Kingdom Egypt which should be interesting.

Is Max Planck still trying to sequence the Al Wusta fingerbone from Arabia? Now that proteome sequencing was successfully used on that Homo Antecessor sample, is anyone looking at using the same technique for the pre-100kya humans/hominids like Skhul & Qafzeh, or Omo?

Man we need a breakthru for PALEOLITHIC DNA in SW Asia/Africa.... all we keep getting are Bronze/Iron Age DNA papers...

100%, but I would also really like to see Upper Paleo stuff from India, Southeast Asia, Oceania, that whole area is a massive black hole when it comes to anything pre-Holocene. East Asia isn't much better but at least we have Tianyuan, and maybe Salkhit (I don't know if that genome has been or will be released at any point?)

[Michalis Moriopoulos]

100%, but I would also really like to see Upper Paleo stuff from India, Southeast Asia, Oceania, that whole area is a massive black hole when it comes to anything pre-Holocene. East Asia isn't much better but at least we have Tianyuan, and maybe Salkhit (I don't know if that genome has been or will be released at any point?)

No kidding. If we get anything from Southern Eurasia and Oceania older than the Hoabinhians, I will be happy. The pre-Holocene East Asian record is still very poor. I'm glad we have Tianyuan, but we need younger samples that show the full East Asian drift. I'm looking forward to seeing what lies between Salkhit (34k BP) and Khaiyrgas 1 (16.6k BP). I'd like to see something purely East Asian in the 25k BP range. Tianyuan and Salkhit are too old, and the latter is admixed with ANE. All the older samples with fully-drifted East Asian ancestry (Khaiyrgas, Ust-Kyakhta) are ANE-admixed, too. If we're going to find a pure East Asian that's over 20 thousand years old, I figure the Amur River Basin or Yellow River area will be our best shot. A sample in that range will be helpful it determing the actual phylogeny of splits between Proto-Amerinds, Northern East Asians, and Southern East Asians. Right now there is conflicting information about whether Proto-Americans split off first (before the Northern and Southern East Asians split off), or if they are just an offshoot of Northern East Asians.

[VladimirTaraskin]

No kidding. If we get anything from Southern Eurasia and Oceania older than the Hoabinhians, I will be happy. The pre-Holocene East Asian record is still very poor. I'm glad we have Tianyuan, but we need younger samples that show the full East Asian drift. I'm looking forward to seeing what lies between Salkhit (34k BP) and Khaiyrgas 1 (16.6k BP). I'd like to see something purely East Asian in the 25k BP range. Tianyuan and Salkhit are too old, and the latter is admixed with ANE. All the older samples with fully-drifted East Asian ancestry (Khaiyrgas, Ust-Kyakhta) are ANE-admixed, too. If we're going to find a pure East Asian that's over 20 thousand years old, I figure the Amur River Basin or Yellow River area will be our best shot. A sample in that range will be helpful it determing the actual phylogeny of splits between Proto-Amerinds, Northern East Asians, and Southern East Asians. Right now there is conflicting information about whether Proto-Americans split off first (before the Northern and Southern East Asians split off), or if they are just an offshoot of Northern East Asians.

Judging by the archaeology of the Amur River, at least in the Russian part, people came there after the LGM, about 18,000 BC. Although there are artifacts in caves of the age of 35,000 BC on the seashore, but this is probably the K2b/P group. Apparently, the peak of the LGM groups N and O were somewhere to the south. Since the Mongolian plateau was probably occupied by the C2a group, N and O were probably even further south, in Middle China. After the peak of the LGM around 18000 BC, group N apparently went north to the Amur River area, group O1 to the south, and O2 remained in the area of the Yellow River.

[Kale]

With two new cool right pops to use and with admixtools2 now, here's revisiting one of my favorite subjects, formation of WHG.
Right pops: 'Iran_GanjDareh_N', 'CHG.SG', 'Vanuatu_ancient', 'Taiwan_Hanben_IA', 'Tianyuan', 'Peru_Laramate_900BP', 'AG3', 'MA1.SG', 'Yana_UP.SG', 'BachoKiro_Gravettian', 'GoyetQ116_1', 'Sunghir.SG', 'Kostenki14', 'BachoKiro_IUP', 'Ust_Ishim.DG'

Spain_Azilian and Spain_Chan_Meso.SG both modeled beautifully as 100% ElMiron (Gravettian added to the list of right pops here).
ElMiron comes out 20.7% Gravettian 79.3% GoyetQ_2
Portugal_Geometric_Mesolithic and Spain_Granada_HG both work as 100% GoyetQ_2 (again with Gravettian in right pops)

There's a pretty tight WHG cluster consisting of Bichon.SG, France_Meso (Rochedane & Ranchot88), France_NouvelleAquitaine_Meso.SG, GBR_Meso, Ireland_Mesolithic.SG, Loschbour.DG, Spain_Canes_Meso.SG
All of them get solid tail probs (between 0.374 and 0.807) except GBR_Meso at 0.065 and Ireland_Mesolithic.SG at 0.046 because both have very low ANE.
Ancestry proportions in the cluster are between 30.0-40.0% ElMiron, 28.0-37.6% Gravettian, and 26.0-33.8% Steppe_Piedmont_Eneolithic.

Our oldest HG Villabruna doesn't come out great either, again because of low ANE, just barely passes, using GoyetQ_2 instead of ElMiron to tighten the std. errors
pop, Piedmont, Gravettian, Goyet, tail prob
Villabruna 32.0 +/- 3.74 51.0 +/- 6.86 17.0 +/- 7.36 0.078

[davit]

100%, but I would also really like to see Upper Paleo stuff from India, Southeast Asia, Oceania, that whole area is a massive black hole when it comes to anything pre-Holocene. East Asia isn't much better but at least we have Tianyuan, and maybe Salkhit (I don't know if that genome has been or will be released at any point?)

The climate there isn't too conducive to DNA preservation. I wonder if we will ever get anything from those regions.

[theplayer]

ancientdnaclimate.png


All "ancient"(includes 100 year old "ancient" Andaman sample...) samples from the "v44.3_HO_public" dataset from David Reich Lab, so ancient dna not on it will be missing, plotted on a Köppen climate map.

(Perhaps I will try to color code the samples by age as well later)

ancientdnaclimateage.png

0-500 BP: white
500-5 000 BP: light red
5 000-15 000 BP: dark red
15 000 - ... BP: black


Edit: just noticed some samples could have the wrong location(for example the 450 BP sample in egypt is supposed to be in south africa, latitude somehow inverted), so might not be 100% accurate

[TuaMan]

The climate there isn't too conducive to DNA preservation. I wonder if we will ever get anything from those regions.

I think the potential game changer for tropical or near-tropical locations will be trying to survey as many caves/rock-shelters as possible in these areas (see Shum Laka in Cameroon) and hope you'll be able to dig out some human remains. If I were the head of a university archaeology department I would mandate all faculty, grad students, and post-docs become semi-professional spelunkers.

Proteomics was also mentioned above, that's another possible work-around. I'm curious though how it compares directly against DNA analysis. With DNA, we can easily tell things like if a person is a Yoruba or Mbuti or a French or a Han or Papuan, or some sort of mix related to any of them. Does proteomics allow for similar granular level analysis as DNA does, or does it only allow for broader inferences?

[davit]

I think the potential game changer for tropical or near-tropical locations will be trying to survey as many caves/rock-shelters as possible in these areas (see Shum Laka in Cameroon) and hope you'll be able to dig out some human remains. If I were the head of a university archaeology department I would mandate all faculty, grad students, and post-docs become semi-professional spelunkers.

Proteomics was also mentioned above, that's another possible work-around. I'm curious though how it compares directly against DNA analysis. With DNA, we can easily tell things like if a person is a Yoruba or Mbuti or a French or a Han or Papuan, or some sort of mix related to any of them. Does proteomics allow for similar granular level analysis as DNA does, or does it only allow for broader inferences?

I didn't think of that. Good point. There is still hope that we get some DNA out of these regions. Either way I hope we are in for a bunch of surprises.

[Kale]

Some more qpadm.
Steppe_Piedmont_Eneolithic works with...
Karelia_HG, Sidelkino, or Arkhangelsk: 29.4-34.5%
AG3: 10.9-14.4%
CHG: 54.5-57.7%
Tail prob: 0.140-0.264
right = c('Anatolia_Epipaleolithic', 'Iran_GanjDareh_N', 'Taforalt', 'Vanuatu_ancient', 'Taiwan_Hanben_IA', 'Tianyuan', 'Peru_Laramate_900BP', 'MA1.SG', 'Yana_UP.SG', 'Italy_GrottaContinenza_Meso.SG', 'Gravettian', 'BachoKiro_Gravettian', 'GoyetQ116_1', 'Sunghir.SG', 'Kostenki14', 'BachoKiro_IUP', 'Ust_Ishim.DG')
Probably overloaded with paleo-Euro pops but meh.

CHG
Iran_TepeAbdulHosein_N.SG: 59.2 +/- 3.19%
Anatolia_Epipaleolithic: 28.6 +/- 4.43%
Karelia_HG: 12.2 +/- 3.13%
Tail prob: 0.419
right = c('Iran_GanjDareh_N', 'Taforalt', 'Vanuatu_ancient', 'Taiwan_Hanben_IA', 'Tianyuan', 'Peru_Laramate_900BP', 'AG3', 'MA1.SG', 'Yana_UP.SG', 'Italy_GrottaContinenza_Meso.SG', 'Gravettian', 'BachoKiro_Gravettian', 'GoyetQ116_1', 'Sunghir.SG', 'Kostenki14', 'BachoKiro_IUP', 'Ust_Ishim.DG')