The deep population history of northern East Asia from the Late Pleistocene to the

[ZetaMaximus]

Ok so I sifted through the raw data on a bunch of samples to look for the mutation, so Yanghshao culture and Qijia culture samples came up carrying non mutated variant but shang dynasty sample has full pair mutation, so looks like the rice farmer thing maybe true. However the other paper thing I was reading mentions theres another mutation on OCA2 in Asians that can cause skin lightening as well, unfortunately I cannot check for it because Reich dataset doesn't have SNP in its dataset.

[Ebizur]

Thank you Ebizur for the citation. I forgot to include it.

The evidence of hpDNA is not to be treated the same as mtDNA, Y-DNA, or auDNA. I like to think of it this way. Today we use spit to determine mtDNA and Y-DNA lines. We use hpDNA to determine "spit lines". H. pylori is transmitted mainly by spit, mainly from parent to child during the weaning phase, before teething allows children to chew their own food. Mostly the mother chews the food, sometimes grandmother if she still has teeth, and in most stone age cultures other close community members can do this too. So it has a matrilineal transmission pattern, but a tendency to homogenise within a community over time. Because of this it is relatively impervious to founder effect. In fact even though it is transmitted most like mtDNA, it has a longterm behaviour closer to auDNA.

I consider studies of the genetic variation and geographic distribution of pathogens like Helicobacter pylori to be of similar import to studies of domesticated plants and animals. They are fun to read and interesting in their own right, but they are not really a valid substitute for human Y-DNA or mtDNA data.

Although bacteria reproduce asexually, they frequently undergo horizontal transfer of DNA resulting in recombination. Therefore, over the long term, as you have mentioned, the pattern of genetic variation for a bacterium such as Helicobacter pylori should resemble that of autosomal DNA: i.e. a smooth cline (except where horizontal transfer of DNA is prevented by some sort of barrier, such as a wide oceanic gap, of which there is certainly no dearth among the Ryukyu Islands -- this is a plausible explanation for the finding of not only one but two different divergent genotypes of Helicobacter pylori isolated from people in Okinawa Prefecture).

Furthermore, there is the additional problem of horizontal transmission of a pathogen among unrelated individuals. Please consider the branch in the dendrogram in Figure 1 of Suzuki et al. 2022 that is comprised of AL05, AL02, AL04, AL03, Aklavik117, Hk840, Hk711, and Hk721. Although these have all been marked with the same dark red color as belonging to "hspAmerind," the three specimens marked with the "Hk" prefix have in fact been isolated from Ainu individuals from Hokkaido, whereas the four specimens bearing the "AL" prefix have been isolated from aboriginal Americans from Alaska, USA. Aklavik is in the far northwest of the Northwest Territories of Canada, about 250 km east of the international border between Alaska, USA and Yukon, Canada. The distributions of mtDNA haplogroup D4 and Y-DNA haplogroup C2-M217 do extend both to Hokkaido and to northern North America, but these haplogroups are very old and widespread. I am not aware of any direct evidence for a particularly close genetic relationship between the Ainu on the one hand and aboriginal peoples of the North American arctic on the other.

[anthrofennica]

I consider studies of the genetic variation and geographic distribution of pathogens like Helicobacter pylori to be of similar import to studies of domesticated plants and animals. They are fun to read and interesting in their own right, but they are not really a valid substitute for human Y-DNA or mtDNA data.

Nor are Y-DNA or mtDNA valid substitutes for hpDNA data! I view them all as complementary. Everything has its place. For linguistics, sometimes domestication studies are more relevant than uniparentals, at least without very comprehensive aDNA.

Although bacteria reproduce asexually, they frequently undergo horizontal transfer of DNA resulting in recombination. Therefore, over the long term, as you have mentioned, the pattern of genetic variation for a bacterium such as Helicobacter pylori should resemble that of autosomal DNA: i.e. a smooth cline (except where horizontal transfer of DNA is prevented by some sort of barrier, such as a wide oceanic gap, of which there is certainly no dearth among the Ryukyu Islands -- this is a plausible explanation for the finding of not only one but two different divergent genotypes of Helicobacter pylori isolated from people in Okinawa Prefecture).

The Ryukyu Islands were "recently" repopulated by the Jomon following the Akahoya eruption. That might rule out a Pre-Jomon origin for the deeply diverged Okinawan clade. The other Okinawan clade is just barely basal to that found in the rest of the East Asians. And this East Asian clade is closer to even the very isolated Sahullans than to the "Deep Okinawans" so we would not expect enough horizontal gene transfer to make the Deep Okinawans closer to East Asians despite waves and waves of East Asian migration. I don't know why exactly the two have yet to homogenise. There are many islands, like you pointed out. But individual population sizes have also been much larger for the past several millenia they were in the Paleolithic. Even the Jomon became sedentary fisher-foragers with surprisingly large individual populations before the Yayoi and Kofun expansions. Which expansion brought with it the East Asian clade? Was one of these at all or something older? We don't have the answer yet. And like you pointed out, conjugation is an issue with hpDNA. But it is also often detectable, thankfully! So it should not pose too severe an issue in the future.

Furthermore, there is the additional problem of horizontal transmission of a pathogen among unrelated individuals. Please consider the branch in the dendrogram in Figure 1 of Suzuki et al. 2022 that is comprised of AL05, AL02, AL04, AL03, Aklavik117, Hk840, Hk711, and Hk721. Although these have all been marked with the same dark red color as belonging to "hspAmerind," the three specimens marked with the "Hk" prefix have in fact been isolated from Ainu individuals from Hokkaido, whereas the four specimens bearing the "AL" prefix have been isolated from aboriginal Americans from Alaska, USA. Aklavik is in the far northwest of the Northwest Territories of Canada, about 250 km east of the international border between Alaska, USA and Yukon, Canada. The distributions of mtDNA haplogroup D4 and Y-DNA haplogroup C2-M217 do extend both to Hokkaido and to northern North America, but these haplogroups are very old and widespread. I am not aware of any direct evidence for a particularly close genetic relationship between the Ainu on the one hand and aboriginal peoples of the North American arctic on the other.

Not sure about a "close" genetic relationship between the Ainu and the Eskaleut, or which proposed relationship it might indicate. But Okhotsk culture left a significant autosomal impact on the Ainu, at about 22% per Sato et al. 2021. And you are no doubt familiar with the Hokkaido medieval and ancient DNA studies on uniparentals. Vovin uncovered a number of linguistic connections between Ainu and Amuric of the sort that one would expect would require a significant fraction of the female population at one point to be L2 speakers. All the North American samples as far as I can tell are Eskaleut speakers (the one from Canada is Inuit, while the "Alaska" samples could be anything unfortunately). Depending on where they came from, they could even be all Inuit. In light of the Chukotkan origin of the Inuit, a closer relationship between the hpDNA of those very specific "Amerinds" (as they label them in the paper) and the South American clade is not very surprising unless perhaps a thorough examination of shared mtDNA clades between the Okhotsk samples and the OBS>…>Thule [and substrata] descendants turns up nothing. And since the relationship in question simply has to be closer to Hokkaido Ainu than to South Americans, we may have a full ~16+ ky (until more is known). I apologise if you have already made such a search for I have neglected to check my notes.

[jose luis]

More like Irula are modeled as 80% AASI/Onge-like and 20% West Eurasian-like, Gujaratis are modeled as 48% AASI/Onge-like, etc. Makes sense that Non-Negrito Orang Asli groups would have >40% East Asian-like ancestry (from Austroasiatic Neolithic rice-farmer groups)

In the study I mentioned the Filipino genomes were almost equally contributed by Papua and the non-indigenous populations in East Asia. On the map from the same study, Basal Asian in Northeast Asia is always below 10%. Missing from this list are the small percentages (up to 3%) of DNA Onge related, in populations of pre-Columbian origin in Latin America. And here, we have ancient DNA (10,400 years old) of the Onge related. Sample Sumidouro5, Southeastern Brazil. Surely, from Southeast Asia, this Basal Asians colonized East Asia till the passage to America in northeast Siberia.

[anthrofennica]

In the study I mentioned the Filipino genomes were almost equally contributed by Papua and the non-indigenous populations in East Asia. On the map from the same study, Basal Asian in Northeast Asia is always below 10%. Missing from this list are the small percentages (up to 3%) of DNA Onge related, in populations of pre-Columbian origin in Latin America. And here, we have ancient DNA (10,400 years old) of the Onge related. Sample Sumidouro5, Southeastern Brazil. Surely, from Southeast Asia, this Basal Asians colonized East Asia till the passage to America in northeast Siberia.

The identity of these "First Americans" or "Population Y" is fascinating to me. Autosomally I have seen them described as "Australasian" or "Australasian-like" in autosomal studies. In what sense exactly I do not know. Do they mean specifically the Australasians, to the exclusion of the South Asians, Andamanese, and East Asians (who were a quadfurcation)?

When we exclude C2a-L1373 (TMRCA 16.2 kya) from the American Y-DNA record, there are still two samples from Brazil ~500 ya (Bot15 and Bot17) within C1b2a-M38 (formed 40.8 kya, YFull TMRCA 18.7 kya but it does not yet include these samples), a branch found at frequency in Melanesia. Columbo et al. 2022 dismissed it as the result of European maritime exchange, but that does not seem very plausible to me, leaving only two possibilities. (1) They discovered an extremely rare Amerind lineage that only managed to survive in a single individual. (2) They discovered a Population Y relict lineage. If the latter is the case, then we ought to pay special attention to C-M38.

Scheinfeldt et al. 2006 gave C1b2a-M38 a TMRCA [42-]49.6[-61] kya, but noted it was rather rare, found in Austronesian speakers of Bougainville although that could have hypothetically come from the Austronesians of New Ireland as it had not yet been found among Papuan speakers (keep in mind however transmission of Austronesian itself in that area was largely matrilineal). They also gave a popular subclade C1b2a1-M208 a TMRCA [39-]46.2[-57] kya.

Mona et al. 2007 focused on the spread of TNG, and found that C-M38 in the Bird's Head region reflected earlier local Y-DNA diversity rather than an intrusion following TNG expansion. In this study we see C-M38 very well-entrenched among Papuan speakers, so I highly doubt it is intrusive in Sahul. In fact it cannot possibly be a recent intrusion, as the C-M38 diversity in the Bird's Head Peninsula area alone reached [6.0-]12.3[-21.3] kya; in Southwest Papua it reached [1.8-]8.0[-16.9] kya. Keep in mind this was still one of the earliest detailed Y-DNA studies of Papuans. This included the domination among the Maybrat (16/24 were C-M38), the Tehit (< West Bird's Head; 5/5 C-M38), Baham (< West Bomberai < TNG; 15/24 C-M38). And a significant presence among the Moi (< West Bird's Head; 2/8 C-M38), Karon (7/22 C-M38, 1/22 specifically C-M208), Moskona=Meninggo (<Mantion-Meax; 5/10 C-M38). There was also 1/24 Ekari (Paniai Lakes < TNG) belonging specifically to C-M208. I cannot remember if it was still this study, but C-M38 was also found in a few Citak (< Asmat-Kamrau < TNG), a few Asmat (< Asmat-Kamrau < TNG), a few Aghu (< Asmat-Kamrau < TNG), and a few Muyu (< Ok < TNG); and specifically C-M208 was found in ~19/20 Dani (< Dani < TNG).

Kayser et al. 2008 looked at Melanesian Y-DNA in the Admiralties, and determined that C-M38>M208 was present at 15%.

Karafet et al. 2010 analysed the Y-DNA of Wallacea, the Philippines, and Sunda. Here they found a very strong presence, but they found it to be strongly associated with Papuan ancestry even on the islands of Wallacea (where they associate it with evidently bidirectional geneflow between Wallacea and Papua), in contrast to some other subclades. They found it in 12/27 Non-Austronesian samples from Alor (< TAP), and in 10/25 Non-Austronesian samples from the Moluccas.

Delfin et al. 2011 focused on the Solomon Islands, including Papuan groups speaking Papuan languages Bilua, Lavukaleve, and Savosavo. They found that including subclades outside the Solomons the haplogroup C-M38 had a TMRCA ~17 kya. They found it specifically in 4/32 Lavokaleve and 1/30 Savosavo; they found the specific subclade C-M208 in 1/44 Bilua. Not very frequent on the Solomons apparently so. To borrow from other studies, we note that the nearest other instance of C-M38 is among the Kuot (2/32), and specifically C-M208 among the Anêm (1/34), and no other instances on the Bismarck Archipelago. So it is only barely found on the Bismarck Archipelago and at low frequency on the Solomons.

Karmin et al. 2015 sampled the Madang at Koinambe/Koinanbe, all 3 of which belonged to a line C-MY1104 (TMRCA ~1 kya, diverged from the Boiken lines).

Bergström et al. 2017 or 2018 (the date got lost from our group's notes) identified a single C-M38 subclade responsible for all Gende (< Madang), of TMRCA [0.8-]0.9[-1.1] kya. This was the same that Malaspinas et al. 2016 detailed, giving them all together a TMRCA of under ~2kya and perhaps closer to ~1 kya (C-MY1096, C-MY1097, C-MY1098, C-MY1099, C-MY1100). In fact this was calculated to be the deepest diverging subclade yet known within C-Y25191 with a TMRCA ~21 kya. So the Gende are a very homogenous Madang group, but still you can see C-M38 must have been on Papua for a very very long time.

Brucato et al. 2021 conducted detailed phylogenetic work on C-M38, finding I believe that it was entirely restricted to Sahul-Wallacea. The Boiken (< East Sepik) belonged half to the branch C-Y25191 (which diverged ~11 kya from its brother clade C-MY1580), half to C-MY1108 (which diverged ~12 kya from its nearest relatives). I think there is even more data on the Boiken from this year, all C-M38 (and at least 2/2 in one group specifically C-M208).

Karmin et al. 2022 studied Y-DNA throughout Oceania. It found the rapid diversification of C-M130 likely to reflect the first settlement of the Sunda and Sahul regions, calculating I think so the split off of most NEA lines except some of the Yaponesian ones at ~51 kya, then ~50 kya for C1-M8 of Japan, then ~49 kya the major breakup. This major breakup was practically a pentfurcation: C8-F725 of {Sunda, Philippines, Flores, Mentawi}, C5-M356 of [mainly West] India, C7-B67 of {Sunda, Flores, Lembata}, C9-B68 of Sunda (specifically Borneo), and the C2-C4 clade. This C2-C4 clade split ~46 kya into C2-M38 of Wallacea and "North Sahul" (TMRCA ~24 kya), and C4-MY1088 of Australia.

It is fascinating to me that a specifically Papuan clade should be found in South America. Even among Australians, so long as the phylogeny they used still holds up (I have not checked ISOGG 2019-2020), studies like Nagle et al. 2016 only found its sister clade C-M347 in Australia, besides basal C. They explained it as a single group of males carrying C-M130 entering Papua, while other related C-M130 descendants migrated further south into Australia (while still part of Sahul). But the situation gets very complicated, both within that paper and in/between others.

For sure today C-M38 is at its highest diversity today on the east side of Papua, but it is still widely distributed, and its deep internal divergences prevent us from fearing it might be a "recent" migration to Sahul-Wallacea. It is imperative to place the samples from Brazil on the C-M38 tree. And in general reconstructing the migrations associated with C are important to the population history of Eurasia and the Americas.

[Pribislav]

One more thing that I might note in this regard is that there are some data (though not much) that suggest that haplogroup C-M8 may be especially common in the Ryukyu Islands.

Hammer et al. (2006) found C-M8 in 4.4% (2/45) of their sample from Okinawa. This figure was actually less than the percentage of C-M8 in their entire set of samples from Japan (14/259 = 5.4%), so Okinawa did not stand out at the time as a likely region of high frequency of haplogroup C-M8.

However, consider the following data points:

Main Island of Okinawa [沖縄本島] (n=36) 16.7% C1 (Totsuka et al. 2016)
Yaeyama [八重山] (n=49) 12.2% C1 (Totsuka et al. 2016)
Okinawa (n=87) 8% C1 (Estimated by Nonaka et al. 2007 based on haplotypes of the Okinawa samples of Rieko Uchihi et al. 2003.)
Miyako [宮古] (n=38) 5.3% C1 (Totsuka et al. 2016)

cf. Rieko Uchihi, Toshimichi Yamamoto, Kanako Usuda, et al., "Haplotype analysis with 14 Y-STR loci using 2 multiplex amplification and typing systems in 2 regional populations in Japan." Int J Legal Med 2003 Feb;117(1):34-8. doi: 10.1007/s00414-002-0319-6. Epub 2003 Jan 10.

cf. I. Nonaka, K. Minaguchi, and N. Takezaki, "Y-chromosomal Binary Haplogroups in the Japanese Population and their Relationship to 16 Y-STR Polymorphisms." Annals of Human Genetics (2007) 71,480–495. doi: 10.1111/j.1469-1809.2006.00343.x.

cf. Shoji Totsuka, The Super Science High School Consortium, Youichi Sato, and Masashi Tanaka, "A study of the geographic distribution of Y chromosomal and mitochondrial DNA haplogroups in Japanese population by Super Science High School Consortium (SSH)." Anthropological Science (Japanese Series) Vol. 124(2), 85–91, 2016. DOI: 10.1537/asj.161018

However, note that Miyako is geographically located between the main island of Okinawa on one side and Yaeyama on the other, and yet the Miyako sample of Totsuka et al. has exhibited the smallest proportion of haplogroup C1. One also should not forget the low frequency of C-M8 in the Okinawa sample of Hammer et al.

Each of the following samples from Japan has exhibited a greater proportion of C-M8 than the Miyako sample of Totsuka et al.:

7/70 = 10.0% Tokushima (Hammer et al. 2006)
4/47 = 8.5% Japanese (Xue et al. 2006)
2/26 = 7.7% Aomori (Hammer et al. 2006)
4/56 = 7.1% Tokyo [JPT] (Poznik et al. 2016)
15/241 = 6.2% Osaka adults (Sato et al. 2014)
6/102 = 5.9% Fukuoka adults (Sato et al. 2014)
22/388 = 5.7% Tokushima students (Sato et al. 2014)
76/1346 = 5.65% Japanese (Mizuno et al. 2010)
18/321 = 5.6% Kawasaki students (Sato et al. 2014)

Therefore, I do not wish to overemphasize this potential association between C-M8 and the Ryukyu Islands, but I do think that it deserves mention.

NAG035, one of the historical (1650-1950 AD) samples from Nagabaka shell midden, Miyako Island, Okinawa, belongs to C-M8, and more specifically to its subclade M8>CTS9336>Z7180>CTS6678>Z7972>Y170131. Speaking of M8, it should also be mentioned that two low coverage samples (TYD006 and TYD007) from Yondaedo, South Korea, dated ~5000 BC, most likely belong to pre-M8.

[Kale]

The identity of these "First Americans" or "Population Y" is fascinating to me. Autosomally I have seen them described as "Australasian" or "Australasian-like" in autosomal studies. In what sense exactly I do not know. Do they mean specifically the Australasians, to the exclusion of the South Asians, Andamanese, and East Asians (who were a quadfurcation)?

When we exclude C2a-L1373 (TMRCA 16.2 kya) from the American Y-DNA record, there are still two samples from Brazil ~500 ya (Bot15 and Bot17) within C1b2a-M38

'Population Y' is described as having elevated affinity to Australians/Papuans, Onge, Tianyuan, populations who really have nothing special in common.
Bot15 and Bot17 if you were just looking at their uniparental and autosomal DNA without knowing where they were from, you'd say 'run-of-the-mill' Polynesian, case closed.
They are almost certainly a case of mislabeling at the museum they were sourced from.
Here's a link to the Botocudo paper.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4370112/

[Ebizur]

NAG035, one of the historical (1650-1950 AD) samples from Nagabaka shell midden, Miyako Island, Okinawa, belongs to C-M8, and more specifically to its subclade M8>CTS9336>Z7180>CTS6678>Z7972>Y170131. Speaking of M8, it should also be mentioned that two low coverage samples (TYD006 and TYD007) from Yondaedo, South Korea, dated ~5000 BC, most likely belong to pre-M8.

Is it possible to identify any private variants shared exclusively between the two archaeological specimens from Yeondaedo vis-à-vis all extant members of haplogroup C-M8?

[Pribislav]

Is it possible to identify any private variants shared exclusively between the two archaeological specimens from Yeondaedo vis-à-vis all extant members of haplogroup C-M8?

Unfortunately no, the coverage of both samples is way too low. Even at M8 level, which has 495 SNPs in YFull, TYD006 and TYD007 have only 3 and 6 SNPs covered, respectively. That should tell you how low is the coverage of these samples.

[Max_H]

That is reminiscent of a qpgraph model I tried a few years ago in which basal Eurasian could be eliminated by having the entire East-Eurasian clade originally forming as a mix of Goyet and some divergent branch, (Basal East Eurasian? lol).

Is been a while since I posted but you said that in this scenario then Ustishim would be more of this Basal East Eurasian ancestry (likely haplogroup D and C for Y lineage and subclade of N and some early M for the mitochondrial lineage) and Tianyuan less so, but both admixed with it? In this case would Hoabinhian be a good proxy for basal East Eurasian or where that ancestry maximizes?