By at least 45,000 years before present, anatomically modern humans had spread across Eurasia [1, 2, 3], but it is not well known how diverse these early populations were and whether they contributed substantially to later people or represent early modern human expansions into Eurasia that left no surviving descendants today. Analyses of genome-wide data from several ancient individuals from Western Eurasia and Siberia have shown that some of these individuals have relationships to present-day Europeans [4, 5] while others did not contribute to present-day Eurasian populations [3, 6]. As contributions from Upper Paleolithic populations in Eastern Eurasia to present-day humans and their relationship to other early Eurasians is not clear, we generated genome-wide data from a 40,000-year-old individual from Tianyuan Cave, China, [1, 7] to study his relationship to ancient and present-day humans. We find that he is more related to present-day and ancient Asians than he is to Europeans, but he shares more alleles with a 35,000-year-old European individual than he shares with other ancient Europeans, indicating that the separation between early Europeans and early Asians was not a single population split. We also find that the Tianyuan individual shares more alleles with some Native American groups in South America than with Native Americans elsewhere, providing further support for population substructure in Asia [8] and suggesting that this persisted from 40,000 years ago until the colonization of the Americas. Our study of the Tianyuan individual highlights the complex migration and subdivision of early human populations in Eurasia.Yang et al., 40,000-Year-Old Individual from Asia Provides Insight into Early Population Structure in Eurasia, Current Biology (2017), https://doi.org/10.1016/j.cub.2017.09.030
Thursday, October 12, 2017
40,000-year-old Tianyuan gives new insights into early population structure in Eurasia (Yang et al. 2017)
Over at Current Biology at this LINK. Here's the summary:
160 comments:
Read the rules before posting.
Comments by people with the nick "Unknown" are no longer allowed.
See also...
New rules for comments
Banned commentators list
As per the main fig, Tianyuan looks on the edge of the ENA clade, however, very basal divergent from the groups that led up to present day East Asians.
ReplyDeleteFor an impression of that, outgroup f3 graphics - Tianyuan vs other early Upper Paleolithic samples - https://imgur.com/a/5l8OQ
(I'm using the outgroup f3 statistics from Davidski's http://eurogenes.blogspot.co.uk/2016/05/on-modern-genetic-affinities-of-ice-age.html, which are the same as in the K14 paper where applicable and the comparable f3 from the supplement here).
Compared to K14 or GoyetQ116-1, Tianyuan shares less ancestry with its closest modern day matches via f3. Slightly more with the closest present day populations than Oase1 and Ust Ishim, though in the case of Oase1 (not Ust Ishim) this is likely to be down weighted by that sample's higher excess Neanderthal ancestry (see figure S3A). Despite modern Europeans having "Basal Eurasian" ancestry (and also speculatively, some ancestry from ENA clade), they still seem closer by outgroup f3 to K14 than East Asians today to Tianyuan.
ENA clade seems likely to me to have a lot of complex structure within it. I'd conjecture it's the ENA who went out east first being replaced later by ENA who stayed west for longer, who contribute more ancestry to East Asians today, and also contributed to Villabruna / Malta clusters (they'd be in the right location to do so, if they were closer to the Near East).
@Matt
ReplyDeleteAs the paper showed Tianyuan to be more related to Suruii than to Mixe it maybe (related to) the mystery population Skoglund saw. Fits your proposition neatly.
@epoch2013, yeah. Their qpGraph sort of resolves that instead by having Surui be mixed between 4% of a population mixed between Papuan related and Tianyuan related population (56:44) and then 96% Native American. I'm a bit suss about that.
ReplyDeleteIn that qpGraph, they do not find a split between First Wave Asian ancestry, that is Papuan Non-Denisovan and Tianyuan clustering together, and then Second Wave Asians - East Asians. Instead Tianyuan and East Asians weakly form a clade.
Though this is very weak - only 2 units away from a total trifurcation!
Could be that if East Asians are a composite with some ancestry from a Tianyuan related pop and another population in the ENA clade, then there was still a First Wave Asian / Second Wave Asian split
The idea of First Wave - through SE Asia - and Second Wave - through Central Asia - ENA waves or just one wave, is something which has gone back and forth in autosomal genetics, basically due to the inconsistent signals of separation found between East Asians and Europeans vs Oceanians.
Two waves has been recently thought to have been made less favourable, mainly by the findings of Denisovan ancestry and one ANE->Siberian+Amerindian ancestry which I think are suggested may resolve these. But we'll have to see - perhaps the pendulum will swing back.
That said, there are versions of the First Wave and Second Wave idea. I think in any case we'd be talking about structure within a single ENA wave, not some of the models where First Wave and Second Wave is the primary split after Basal Eurasian...
Hope there are enough bones out there in the world to actually resolve a lot of this.
I was just having a discussion with Andrew over at his blog about substructure in ENA which may be relevant here.
ReplyDeleteExtra Oceanian ancestry for some populations in the Americas. The Surui are much closer to the Onge than the Mixe are.
https://www.nature.com/articles/nature14895.epdf?referrer_access_token=jLOJcw9ZkEUkitvKpb4icdRgN0jAjWel9jnR3ZoTv0N6yB-nEyCdRoL51ykMO5E91orKw8mXMzuxoORJJBQhEhMryvaarH9hFKiGoJjc1JvX7zNtkTMbwuOPicFdsM1y7uzyyWsvTeUZY4iN3Kq60ZhvW781KSFuU4fHlV0EhySXsodIVFx76RjUcfg3OxAfFtAfWeS579uPXTb94CfsRhYDaIFWuW50EJJArdUBaLgEMHNKjN0PA9CwAMcJjrQ0&tracking_referrer=www.nature.com
Denisovan ancestry in the Americas: https://academic.oup.com/mbe/article/32/10/2665/1210305
It seems that populations with excess Oceanian ancestry are mutually exclusive with those that have excess Denisovan ancestry. I have to go visit my mother in the hospital now so I can't check myself, but does it look like Tianyuan has any special relationship to either of these indigenous American groups? IE does excess Tianyuan ancestry in the Americas correlate with excess Oceanians, or with excess Denisovan, or neither?
How about South Asia? There's an excess of Denisovan in some populations there too that is unexplained. Does there seem to be a similar pattern with this ancient sample?
@ Matt
ReplyDeleteInteresting points.
*"Compared to K14 or GoyetQ116-1, Tianyuan shares less ancestry with its closest modern day matches via f3."
Nevertheless, they argue 'our results indicate that the Tianyuan individual is related to an ancestral group that contributed to all more recent populations with Asian ancestry'
However, he is *not directly* ancestral, and Tinyuan branches off prior to the population which would be directly ancestral to present day East and SEAs.
* On the TY <-> Goyet Q116-1 connection.
So far restricted to them, and "ancient European individuals related to GoyetQ116-1,
such as the 19,000-year-old ElMiron individual from Spain [4], do not share more alleles with the Tianyuan individual than other ancient European individuals do (Figure S2B). Present-day East and Southeast Asian populations do not share significantly more alleles with GoyetQ116-1 than they do with other ancient Western Eurasians (Figures 2C and S2A).'
That fits in with point above.
What it suggests is that Tinyuan has deep links to GoyetQ116-1, somehow, and also certain modern South Americans. We're looking at a distinct, yet unsampled population, which contributed to both Tinyuan and Goyet Q116-1 ? They hint at the mtDNA Hg M connection.
As Matt suggests, this was then largely swept over by a newer wave of 'newer ENA', which was expected even from pre-aDNA days.
Additionally, this newer ENA is also that which is see in La brana, so the 'eastern' links in Goyet & La Brana come from different times / groups.
This might be quite interesting as well, though hard to interpret:
ReplyDeleteFig S2 D shows that the Villabruna cluster are approximately equally related to Tianyuan as Vestonice16... but are more related to later Asians than Vestonice16 is (most so for Loschbour and least so for Villabruna: https://imgur.com/F5N2Td5
The above shows the D scores, but you can see the Z in Table S3 iii): https://i.imgur.com/8FMMwVV.png
(essentially around zero for the comparisons D(X, Vestonice16; Tianyuan, Mbuti) and divergent for D(X, Vestonice16; Recent ENA, Mbuti)).
D statistics in the supplement Table S3a Z3 also show that Tianyuan is a) approximately equally related to K14 and Villabruna cluster (very slightly more related to VB cluster), b) less related to CHG and EEF than to K14, c) more related to GoyetQ116-1 and Mal'ta1 than to K14: https://imgur.com/lG3fVnx
(Just gave this a very quick skim, and haven't looked at the supplementary material; I'll do a proper "reading", eventually)
ReplyDeleteOkay, so GoyetQ116-1 also had ENA admixture (as did ANE, EHG, and most samples within the broader "Villabruna" cluster).
Also, @Matt:
Thanks for the outgroup f3 graphics; very interesting stuff.
"Taken together, our results indicate that the Tianyuan individual is related to an ancestral group that contributed to all more recent populations with Asian ancestry. Also, the Tianyuan individual’s age indicates that a genetic separation of Europe and Asia must have been earlier than 40,000 years ago. This is consistent with a split time of 40,000–80,000 years ago estimated for European and Asian populations based on mutation rates estimated from de novo mutations."
ReplyDeleteThe only population on all of their TreeMIx charts that is connected to both MA-1 (upstream of all ancient and modern Eurasians, as we have abundantly seen recently) and Tianyuan (upstream of all ancient and modern East Asians) are Amerindians.
It stands to reason that "a split ... of 40,000–80,000 years ago estimated for European and Asian populations" must have happened in America. Out-of-America is beautifully confirmed again...
@German
ReplyDeleteOut-of-America is beautifully confirmed again.
There's absolutely nothing in this paper that even remotely hints at Out-of-America.
But hey, why don't you float your unique interpretation of the data and models in this paper to the authors, and let us know if you even get a reply.
@Davidski
ReplyDeleteAbsolutely nothing but the data hints at out of America.
What "reply" are you talking about? I'm giving answers, not asking questions.
@German
ReplyDeleteYou're seeing something that no one but a few random crackpots online would ever admit to seeing. Well done if you're right. Go and lay it out in a brilliant post at your blog.
And if you are right, then expect phone calls from Nature, Science, etc. sooner or later, and an apology from me. But to be honest, I don't think anyone will ever call.
@Davidski
ReplyDeleteI couldn't care less what you think about anything beyond Bronze Age Europe. Why would they call me? It's like expecting the Pope to call Richard Dawkins once Dawkins "gets it right." Do you think Dawkins has ever been waiting for a "reply"?
Interestingly in the Treemix here the Chokhopani burials split just after Tianyuan does, and before Native Americans receive their ancestry from the main East Asian clade. This contradicts the Sherpa and Ainu papers where Native American ENA splits before Ainu and Highland Tibetan ancestry does. Someone really should get Chokhopani/Sherpa and Jomon/Ainu genomes and run them together with Amerinds, Tianyuan, Papuan and Onge to get a clear phylogeny of structure in the presently known branches of ENA.
ReplyDelete@German
ReplyDeleteNo idea if Dawkins waits for replies, but I'm pretty sure he gets calls.
@davidski
ReplyDeleteSure he does, but not from the popes of the world
Everybody knows Richard Dawkins doesn't receive phone calls, he calls them. Yet, I'm sure Francis Collins gets these phone calls.
ReplyDeleteWhy they didn't tell what is his Y-DNA? The sample is male and the coverage is sufficient.
ReplyDelete@Seinundzeit - "Okay, so GoyetQ116-1 also had ENA admixture (as did ANE, EHG, and most samples within the broader "Villabruna" cluster)."
ReplyDeleteWe probably should stop calling it ENA if that seems to not be a valid clade anymore.
I think the mtDNA M thing is probably right and maybe from an older migration out of Africa. Maybe that's true of Y haplogroups C and D too?
This would actually explain a lot. For example, on the Y-DNA side, East Asian Y-DNA seems to be nested within West Eurasian Y-DNA - all haplogroup F descendants exclusively West Eurasian except K2, and K2's descendants are all deeply rooted in East Eurasia. This is exactly what we would expect if humans with haplogroup F migrated into West Eurasia before a subgroup later moving on to East Eurasia with K2.
F is in turn nested within a deeper layer of East Eurasian Y-DNA variation though. All of the descendants of CT are exclusive to East Eurasia, with just two exceptions:
F, which is exclusive to West Eurasia
E, which seems to be only recently arrived in West Eurasia, and seem to originate from Africa's closest point to East Eurasia.
This is exactly what we would expect if either there was
So I think this fits pretty nicely with two migrations to Eurasia:
Migration 1, carrying Y-haplogroup CT and mtDNA M or maybe just L3. This spread throughout the whole of Eurasia more or less simultaneously.
Migration 2, carrying Y-haplogroup F and mtDNA N. This descended from a subgroup migration 1, and spread first from West Eurasia to the whole of Eurasia.
I'm hesitant to call either an "Out-of-Africa" migration though because one or both could have from a "basal Eurasian" group geographically in Africa, or they could have spread from Arabia, and migration 1 could have spread from India or SE Asia.
I believe Dienekes has been suggesting something like this for a while too?
@German - if you think your model is a better fit then do it and show us the goodness of fit. Don't just say that your model is better - prove it.
And if you can't, then please just move on from this Out-of-America nonsense. Science means testing your hypothesis. There are all the tools available to do so. Either it's supported or it's falsified.
@Ryan
ReplyDelete"if you think your model is a better fit then do it and show us the goodness of fit. Don't just say that your model is better - prove it.
And if you can't, then please just move on from this Out-of-America nonsense. Science means testing your hypothesis. There are all the tools available to do so. Either it's supported or it's falsified."
You just copy pasted this from a smart site of sorts, didn't you? "prove it" "falsify" "test". How in the world is out of-Africa that you just mentioned "proven" or "tested"? It was falsified from the very beginning. Nothing genetically African has ever been discovered outside of Africa until slave trade times. You can see those African genes of massive antiquity all over America since 1492 but you don't see them in any ancient Pleistocene DNAs in Europe, Asia or America. What are building your ideas out of of? Thin air? Just like the pseudohistorical "Basal European" and "ANE" constructs?
@ German Dziebel
ReplyDelete“@Davidski
Absolutely nothing but the data hints at out of America.
What "reply" are you talking about? I'm giving answers, not asking questions”.
You know I am grateful to Davidski for many reasons, but your answer delights me, because we are “also” men of letters beyond the scientific part, you of course gets larger than me.
See Picasso (whose surname from the maternal side came from Liguria, Italy):
I don't search, I find... or similar. In Italian : Io non cerco, trovo.
@Ryan
ReplyDeleteNo amount of data and testing will change the mind of a creationist. He will always see God's work in Dilophosauruses. The same with human origins: academic scholars and their web groupies will always see out-of-Africa where nothing of this sort appears to a naked, unbiased eye. You can easily detect Neanderthal DNA, Denisovan DNA outside of Africa (and in Africa if you look better), but there are no African genes outside of Africa.
@ German Dziebel
ReplyDelete“but there are no African genes outside of Africa”
We have even hg. A and B out of Africa, but no A00, which is at the level of the Neanderthals and Denisovans and many others very likely.
How to explain that the oldest splitted hg. L is in Sardinia and not in India?
@Gioiello
ReplyDeleteThere were either carried out of Africa in relatively recent times or if they are indeed basal, then it shows that L may have formed in close geographic proximity to Africa, i.e. the Middle East, Southern Europe..., and then expanded into Africa. An alternative would be to propose that mtDNA L, Y-DNA A, B were absorbed from archaic humans in Africa. We can debate those options but we should give out of Africa to the poor.
@German - "but there are no African genes outside of Africa"
ReplyDeleteThat's circular logic nonsense. You assume that there's no African genes outside of Africa because you assume that no basal genes are African. 100% of non-Africans are y-haplogroup A. 100% of non-Africans are mtDNA haplogroup L. Those are both African haplogroups.
"No amount of data and testing will change the mind of a creationist."
So you won't bother ever looking at data then? Data won't convince people therefor data has no value? Why not just admit that data doesn't support this.
@Ryan
ReplyDelete"That's circular logic nonsense. You assume that there's no African genes outside of Africa because you assume that no basal genes are African. 100% of non-Africans are y-haplogroup A. 100% of non-Africans are mtDNA haplogroup L. Those are both African haplogroups."
No, you are again using words the meaning of which escapes you. Out-of-Africa is based on circular logic. Academic scientists produce the data, but they don't really think through it. Those are different skill sets. What's the best way to explain it... Here you go. Try to grasp it: Artists are no art historians. Artists produce works of art, art historians makes sense of it.
The way to go beyond circular logic and to look at the data. You can clearly see Y-DNA A00 in African Americans but you don't see it in Ust-Ishim. Slave trade happened, while out of Africa did not. How old are you, Ryan?
I wrote that the language spoken of A00 hasn’t arrived till us and all the African languages came from outside. Neither the languages of hgs A and B arrived till us. You of course have wide knowledge in many fields and you compared the numerous American phyla to the few African ones.
ReplyDeleteBut if I had the programs and the knowledge I’d want to check the Shi Huang theory.
@Ryan
ReplyDeleteWith Amerindians it's all very straightforward: unlike Africans whose genes are not reflected in aDNA from Eurasia, aDNA that's basal to West Eurasians (MA-1) and to East Asians (Tianyuan) has strong affinities to modern Amerindians but not as much to, respectively, East Asians and West Eurasians. It's pretty clear that if West Eurasians and East Asians split from a parental population (which everybody agrees on), this parental population likely lived in America.
@Matt
ReplyDelete"This might be quite interesting as well, though hard to interpret:
Fig S2 D shows that the Villabruna cluster are approximately equally related to Tianyuan as Vestonice16... but are more related to later Asians than Vestonice16 is (most so for Loschbour and least so for Villabruna: https://imgur.com/F5N2Td5"
Confirming that the Villabruna gene flow is something relatively recent in both Villabruna, likely ANE as it is consistently accompanied with American affinity, something Goyet hasn't.
Could Goyet be contaminated?
@German - Show it. Test the model. Show us that the model is a better fit than theirs. I'm looking for hard data not words.
ReplyDeleteGerman is basically xyyman with better punctuation. Don't feed the troll.
ReplyDelete@ Matt
ReplyDeleteI wonder if the recently published Mesolithic from Norway also has the more recent eastern admix ?
If so then there is a route- far NEE (Karelia) to KO1, on one hand, then to Western Europe and down Atlantic to LaBrana
German Dziebel is good example of an articulate and apparently educated person who says stupid things. Often people like him are also douche bags. When they realize they're totally wrong they resort to insulting their opponents' intelligence rather than engaging in a legitimate debate.
ReplyDeleteAlso, remember lots, not most, of DNA papers based on modern DNA, make stupid theories. Publication=/=Correct.
@Samuel - to be fair I'm the one being the dick here, not German. I think he's dead wrong but I haven't seen him be a jackass to anyone that I can recall. Seems like a decent enough guy.
ReplyDeleteInteresting that Mal'ta split from the Western Branch very early....
ReplyDeleteCapra: You nailed it with that comparison.
ReplyDelete@ Epoch2013, check out the QUANTIFICATION AND STATISTICAL ANALYSIS, which actually has a good set of discussion of the phenomena I linked in my comment there (which I probably should have read and noticed before trying to reinvent the wheel):
"Comparisons of Asians to more recent West and North Eurasians
------------------------------------------------------------------------
Comparing ancient North (ANE = Mal’ta1, AfontovaGora3) and <14 kya West (EUR<14 kya = Villabruna, Bichon, Rochedane, Ranchot88, Loschbour, LaBrana1, Hungarian.KO1, Karelia and Motala12) Eurasians to EAS and the Tianyuan individual, as well as other Eurasians (EUR = Kostenki14, Vestonice16, ElMiron, Villabruna), we find the ANE and EUR<14 kya form a clade with the EUR (Tables S3Di, S3Dii, S3Ei, and S3Eii), but D(EUR, ANE/some EUR<14 kya; EAS, Mbuti)<0 (Tables S3Diii and S3Eiii; Figures S2C and S2D), suggesting a connection between EAS and ANE, and EAS and some EUR<14 kya.
We refer to the Bichon, Loschbour, LaBrana1, Hungarian.KO1 as the VWE, and Villabruna, Rochedane and Ranchot88 as the VNE in further analyses.
The VNE do not tend toshow a connection to EAS, whereas the VWE do, shown in Table S3Eiii and in [4].
The ANE, Motala12 and Karelia show a connection to the EAS, but a stronger connection to the AMER, while the VWE show a similar connection to both the EAS and AMER (Figures S2C and S2D).
The Tianyuan individual may show a slight connection to the ANE (Table S3Diii; Figure S2C), but the single significant result, D(Kostenki14, Mal’ta1, TY, Mbuti)<0, is not significant when using transversions only.
No connection is observed for the VWE (Table S3Eiii; Figure S2D).
Lazaridis et al. [16] argue that West European hunter-gatherers (WHG: Loschbour, Hungarian.KO1, and LaBrana1) can be considered a mixture of populations related to East European hunter-gatherers (EHG: Karelia) and an older hunter-gatherer from Switzerland (Bichon, SI7 of Lazaridis et al. [16]).
Here, we ask whether the Asian signal we observe in the VWE (composed of the WHG and Bichon) is related to the EHG’s connection to EAS. First, [4] and we find that the Bichon individual also shares a connection to EAS. Second, for Karelia, we observe larger D values for AMER than for EAS, (Figure S2D; Table S3Eiii), likely related to the influence of ANE on Karelia [16, 20]. This difference is not observed in the WHG (Figure S2D). In contrast, Motala12 shows similar results to the WHG but does have elevated D values when compared to AMER (Figure S2D), suggesting Motala12 is connected to Karelia or the ANE.
The Asian connection in the VWE cannot be explained by a connection to Karelia or the ANE.
Using D(EUR<14 kya, Kostenki14/GoyetQ116-1, TY, EAS) (Table S3C), we show that EAS share more alleles with EUR<14 kya than the Tianyuan individual shares with them, relative to Kostenki14 and GoyetQ116-1, further supporting connections between EAS and ANE, the VWE, Karelia and Motala12. We note that EAS also share a connection to Satsurblia, Stuttgart, and the French and Sardinian, indicating other potential interactions unrelated to ANE or VWE.
E.g. they are more or less saying that VWE samples are closer to the ENA part of ancestry in both AMER and EAS, as they are equally shifted towards both relative to Vestonice, and that the VWE appear to lack ANE ancestry, as they have no shift of either a) greater to AMER than EAS or b) towards AMER relative to what Vestonice has, after the general EAS shift is considered.
At the same time, VWE are not shifted toward Tianyuan relative to either VNE or Upper Paleolithic Vestonice.
@epoch,
ReplyDeleteRe; contamination of GoyetQ116-1, I can only really quote this from the same section:
To check for robustness of the GoyetQ116-1 result, we test for shared errors in GoyetQ116-1 and the Tianyuan individual related to ancient DNA (aDNA) damage, DNA contamination, or data processing, rather than past demographic processes. We find similar results when we restrict to only transversions (jZj>2.2, D value remains similar; Table S3Bii), to avoid sites where differences might be errors from cytosine deamination due to ancient DNA damage rather than an actual allele from the ancient individual.
Restricting to damaged fragments reduces possible present-day DNA contamination, as it only merges sequenced DNA fragments showing characteristics of aDNA damage. Tianyuan libraries were double-stranded and uracil-DNA-glycosylase (‘ds UDG’) treated, meaning damage restriction could not be performed, but GoyetQ116-1 libraries were either single-stranded with UDG (‘ss UDG’) treatment or double-stranded without UDG (‘ds noUDG’) treatment, so damage restriction could be performed (Box 2.1 in SI2 of Fu et al. [4]). The Tianyuan/GoyetQ116-1 connection is still found using damage-restricted fragments (Table S3Biii; Figure S2Aiii). We also find D(Stuttgart, French/Sardinian; TY, Mbuti)0 (Z = 2.8/0.8) and we observe a weaker relationship between EAS and GoyetQ116-1 than between the Tianyuan individual and GoyetQ116-1, relative to other ancient West Eurasians (Figures 2C and S2A), so results are not consistent with present-day contamination.
All ancient individuals were similarly processed, but only GoyetQ116-1 shows the Tianyuan connection, making it unlikely the signal is an artifact of data processing. Our results were robust to outgroup choice except when using archaic individuals (Tables S3Bi–S3Biii). However, slight differences in how GoyetQ116-1 and the Tianyuan individual relate to these archaic individuals may mask the signal (Table S4A). Shared higher archaic admixture proportion in the Tianyuan individual and GoyetQ116-1 cannot explain the signal, as we do not observe that the Denisovan or Altai Neanderthal share more alleles with the Tianyuan individual and GoyetQ116-1 than with other ancient Eurasians of a comparable age (Tables S4A and S4B). No correlations in sequencing error rate, reference bias, or GC-content bias for GoyetQ116-1 or the Tianyuan individual can explain why the connection is only observed for these two individuals (Table S3G).
Hard to think why contamination of modern->ancient could result in any very specific patterns of similarity between these two ancients.
Does this have something to do with something more archaic than Denisovans, Neanderthals and AMH ? Maybe that unknown species talked about ? Maybe something closer to Heidelbergensis....
ReplyDeleteAnother thing to note which everyone will have already picked up on is there is an inconsistency between the main qpGraph and the base model figure which opens the paper. The final qpGraph has admix into MA-1 from a population *upstream* of the split between all ENA, while the base model figure has the ENA to MA-1 edge *downstream* of the split between Tianyuan and EAS.
ReplyDeleteKind of hard to see how a downstream split works with fairly pronounced shifts in relatedness to EAS vs Tianyuan in Karelia. However, perhaps MA-1 does not share this same pattern.
E.g. the table S3Ci shows the following:
D(Malta, Kostenki14; Tianyuan, Ami) = Z: -0.7
D(Malta, Kostenki14; Tianyuan, Miao) = Z: -0.9
D(Karelia_HG, Kostenki14; Tianyuan, Ami) = Z: -2.3
D(Karelia_HG, Kostenki14; Tianyuan, Miao) = Z: -3.2
D(Loschbour, Kostenki14; Tianyuan, Ami) = Z: -2.1
D(Loschbour, Kostenki14; Tianyuan, Miao) = Z: -2.7
D(Satsurblia, Kostenki14; Tianyuan, Ami) = Z: -3.8
D(Satsurblia, Kostenki14; Tianyuan, Miao) = Z: -3.8
D(Stuttgart, Kostenki14; Tianyuan, Ami) = Z: -3.9
D(Stuttgart, Kostenki14; Tianyuan, Miao) = Z: -3.7
(Btw, attraction between EEF/CHG and EAS relative to Tianyuan couldn't be because of Basal Eurasian - in theory BEu should reduce relatedness to both TN and EAS... Hopefully this is not just some artefact).
Still looking over the alternative qpGraphs in the supplement and the corresponding section "Testing a model including the Tianyuan individual".
This includes the note: "We also test adding GoyetQ116-1 or the Onge to the final model (Figure S4N), but could not find a graph that fits the data, suggesting complex demographic parameters outside the parameter space of AdmixtureGraph (i.e., population structure, continuous migration, etc.) may play a role. However, we note that the best-fitting graphs (Z = -3.5) always place GoyetQ116-1 as a mixture of a population related to Europeans and the Tianyuan individual. The best-fitting graphs for the Onge (Z = - 3.4) always include them as a mixture of populations related to the Ami and Papuan."
@Rob, I'd expect so, though at some point in the future someone could certainly test all the Mathieson, Gunther, Mittnick, González-Fortes, etc. 2017 Mesolithic-Holocene Euro HG samples.
I can't seem to find any Y chromosome reads in the BAM file, even though the study says the Tianyuan individual is male.
ReplyDeleteAnyone have any idea why, if the coverage is 2.98x?
This comment has been removed by the author.
ReplyDeleteMaybe Haplogroup A ? Heheheeh....
ReplyDelete@Samuel Andrews
ReplyDeleteLet me "insult" your intelligence as well: Ryan or you are not my "opponent" just like a patient is not the opponent of a doctor. I'm just telling him what's going on. He has an option of going with a professional opinion or keep self-medicating.
@Capra
"German is basically xyyman with better punctuation."
With people like you around, xyyman is perfectly at home at this blog. He never shows up on mine.
@OG, in the excel sheet it states that the amount of the Tianyuan individual's nuclear data available; the average coverage ranging from panel: panel 1-3 IS 4.09, panel 4 is 1.71, and panel 1-4 is 2.98. All of this is described in S6. So, I guess your correct the average coverage is 2.98.
ReplyDeleteSince we are diploid organisms, it is preferable to use diploid aDNA for downstream analysis, as that would more accurately depict the original genomes than the pseudo-haploids that are used in the overwhelming majority of cases.
ReplyDeleteWhile it is true based on my own work that at low coverages, <2X, the diploids will start converging with haploids in terms of accuracy as defined by hetrozygosity, reference bias, low frequency variants, and other metrics, because a large proportion of the bases are supported by only 1 read, for coverages >3X, however, diploids convey a more accurate picture, especially in haplotype and IBD comparisons, according to my work, the results of which will be available at EurasianDNA.com in the next couple of weeks.
Although the differences are not as noticeable using ADMIXTOOLS, because accuracy on a single strand is adequate for good results, nonetheless, here are diploid-haploid dstats for Srubnaya I0232, which is 8X coverage. The diploids I processed.
For example, notice how the diploid one share more drift with modern pops which are more proximate to the sample's recovery location (Ukranians, Tatars, Kurd), whereas the pseudo-haploid one shares more drift with NW Europeans.
Srubna_I0232_Diploid Srubna_I0232_Haploid Kurd_C Chimp 0.0107 1.543 107065
Srubna_I0232_Diploid Srubna_I0232_Haploid Ukrainians_east Chimp 0.0068 1.13 105599
Srubna_I0232_Diploid Srubna_I0232_Haploid Ukrainians_west Chimp 0.0089 1.443 105599
Srubna_I0232_Diploid Srubna_I0232_Haploid Tatars Chimp 0.0083 1.343 105599
Srubna_I0232_Diploid Srubna_I0232_Haploid Icelandic Chimp -0.0041 -1.277 388213
Srubna_I0232_Diploid Srubna_I0232_Haploid English Chimp -0.0033 -1.039 388213
Srubna_I0232_Diploid Srubna_I0232_Haploid French Chimp -0.0036 -1.108 388213
Modern human contamination from fossil handlers can also alter aDNA genotypes, proportional to the amount of contamination, and it is difficult to address. Although I have analyzed this in some of the steppe samples, I have no idea how much of a factor it is in the ones being discussed here.
@Synome
ReplyDeleteOne would think this sample is too old to raise ethnic sensitivity problems with its Y-DNA. Also is this something that ever happened? It is too early for conspiracy theories.
@Ryan
If I remember correctly, Dienekes suggested that the main source of the 60 kya Eurasian expansion was a population in SW Asia/Arabia that actually came out of Africa much earlier, during the last interglacial 120-130 kya, but did not have much lasting impact farther in Eurasia at first attempt. He did not elaborate on the supposed movements in Asia, but he wrote this in a time, when the aDNA data were orders of magnitude less, and practically nonexistent in Asia.
For those having a hard time interpreting what I posted above, one of the things it shows is that the actual Srubna individuals (also Yamna and Andronovo) were a little more “Asian” and a little less “European” than what analysis using the published pseudo-haploids show. This is consistent with their recovery locations
ReplyDelete@Matt
ReplyDeleteThanks a lot Matt.
@Matt
ReplyDeleteAccording to Fu et al the following samples can be sampled as Villabruna + Goyet: La Brana, Loschbour and Ranchot88, while Bichon and Rochedane could be modeled as pure Villabruna. That doesn't match with VNE/VWE distinction you pasted
There goes the theory that this affinity was brought with Goyet. It is not distinctly separated geographically as Bichon apparently has it and Ranchot (Jura, nearby) hasn't. Not temporarily separated as Villabruna is comtemporary to Villabruna hasn't. Hell, Villabruna is even closer to east Asia..
I can't make anything of this.
@Matt
ReplyDelete"Villabruna is contemporary to [i]Bichon[/i]" that should have been.
@Kurd
ReplyDelete"For those having a hard time interpreting what I posted above, one of the things it shows is that the actual Srubna individuals (also Yamna and Andronovo) were a little more “Asian” and a little less “European” than what analysis using the published pseudo-haploids show. This is consistent with their recovery locations"
http://eurogenes.blogspot.com/2017/08/a-bronze-age-dominion-from-atlantic-to.html
Geography =/= Genetics
"We find that he is more related to present-day and ancient Asians than he is to Europeans, but he shares more alleles with a 35,000-year-old European individual than he shares with other ancient Europeans, indicating that the separation between early Europeans and early Asians was not a single population split."
ReplyDeleteThat argument does not prima facie logically follow. Maybe there is a reason esoteric to the layman? I am a casual layman in genetic anthropology but I fancy that am actually quite good at logic.
@German Dziebal,
ReplyDelete"Let me "insult" your intelligence as well:"
Funny you put insult in quotes. Countless posters here say you insult them. The only person who dis agrees is yourself. When the guilty claims innocence and countless viewers claim he's guity he's usually guilty.
German, you're a typical arrogant "intellectual" snob who is disturbed because despite your pride and education an average joe has whooped your butty in a debate. You're of the same kind as Rob, every Liberal Night Show Host, and half the moderators at Anthrogencia.
"Let me "insult" your intelligence as well: Ryan or you are not my "opponent" just like a patient is not the opponent of a doctor. I'm just telling him what's going on."
What is it you are trying to inform me of? B.S. like that mtDNA recombines like autosomal DNA does. Buy an mtDNA for Dummies Book, man.
mtDNA Doesn't Recombine doesn't happen!! The same mutation rarelly ever appears in to seperate maternal lineages!!!
I am familiar with these facts because I have personally examined 10,000s of mtDNA samples. The currently accepted human mtDNA phylogenetic tree posted at Phylotree.org is correct.
http://phylotree.org/tree/index.htm
Eurasian mtDNA is a subset of African mtDNA diversity. Ameridian mtDNA is a subset of Asian mtDNA diversity. Neanderthal mtDNA and Eurasian mtDNA don't match. There's no ifs, ands, or buts about these things. For it not to be true literally like 50 mutations would have to appear in seperate mtDNA lineages. And that's impossible!!
German, you're abosultly full of poop. No one here agrees what you say. I've come up with some explinations for why you post BS and act like an asshole.
-You're insane/mentally ill.
Oh, wait that's only one explination. Yeah, I think you're an insane freak. If you were a nice crazy person I'd be nice back but you're a giant douche bag.
@Samuel Andrews
ReplyDeleteI can see how eating a trillion mtDNAs for breakfast made you feel special but not to me. You are not my opponent, you are an object of my anthropological interest. You are about as special to me as a broken pipe to a plumber. I don't care if people disagree with me or they feel insulted - believers will always "disagree" with scientists and will always feel that scientists are mocking them. In reality, I've never called anybody a "douche bag" or anything of that nature. I just advanced a theory and it's being proven right with every aDNA study.
Y-DNA and mtDNA trees are correct? So says, with confidence, one Sam Adams. But the truth is, there are no African-specific lineages found anywhere in aDNA samples in Eurasia. Just look at the distribution of African lineages in the New World post-1492. You can see what an out-of-Africa migration looks like. Now try to find the same pattern in Eurasia. Nothing of that sort. This means that 1) either they got absorbed from African archaics - that's why they never exited Africa; or 2) they are fully modern lineages but phylogenetically they sit underneath either mhg M or mhg N. Shi Huang argued that African L lineages derive from Eurasian M lineages. This would be the best scenario that will explain everything. We will see. But don't start squealing that Shi Huang is an agent of Chinese imperialism. Out-of-Africa is just as much as product of Western liberalism as out-of-Asia is a product of Chinese imperialism.
@German,
ReplyDelete" I don't care if people disagree with me or they feel insulted - believers will always "disagree" with scientists and will always feel that scientists are mocking them."
Lol. Bro, it's the other way around. ALl of us who dis agree with follow science. You follow a bogus theory you came up with years ago which has been proven wrong but you refuse to let go.
"But the truth is, there are no African-specific lineages found anywhere in aDNA samples in Eurasia. Just look at the distribution of African lineages in the New World post-1492. You can see what an out-of-Africa migration looks like."
I and the rest of the genetic community refer to a different kind of African. We are not referring to recent Africans we're referring to humans who lived in 60,000+ years ago.
The modern day African mtDNA gene pool did not exist 60,000 years ago. I guarantee you the vast majority of widespread, African-specific lineages today emerged less than 20,000 years ago.
That's the case for Europe and the Middle East. Most lineages that connect people from Iran to Ireland emerged less than 20,000 years ago (eg, T2, J1b, H, K1a). Within Europe the vast majority of Europe-wide lineages formed 10,000 or less years ago (T2b, H1, K1a4a1, U5a1a1, etc).
The point is, modern day mtDNA gene pools formed long after 60,000 years ago. There's a 0% chance Africa's mtDNA gene pool looked like it does today 60,000 years ago. Therefore the mtDNA impact of the forced migration of Africans to America in recent times is not comparable to the mtDNA impact of the migration of Africans to Asia 60,000 years ago.
One, shouldn't expect to find young African specific lineages from an African migration that occured 60,000+ years ago. L3, the mother of Eurasian M and N, is estimated to be around 80,000 years ago. Most of its subclades are found in Africa not Eurasia. All of L3's relatives: L4, L2, L1, etc, are found in Africa NOT Eurasia. All of that strongly indicates L3, the mother of Eurasian mtDNA, orignated in Africa.
When a migration from one region to another occurs not every mtDNA lineage takes part in the migration. FOr example, not every mtDNA lineage that existed in the Neolithic Near East moved into Europe. Only some did. That's why most mHG H, T, J, and K lineages in Europe are rarelly found in the Middle East today even though they orignated in the Middle East 8,000+ years ago.
@German,
"This means that 1) either they got absorbed from African archaics - that's why they never exited Africa; or 2) they are fully modern lineages but phylogenetically they sit underneath either mhg M or mhg N. Shi Huang argued that African L lineages derive from Eurasian M lineages."
Omg, you are such an idiot. How many times do people have to tell you: The current phylogenetic tree is correct. Recurring mutations can not explain the currently accepted tree.
Eurasian mtDNA is a subset of AFrican mtDNA diversity. There's no ifs ands or buts about this!!
Do you understand that if the currently accepted mtDNA tree is correct, which it absoloutly is, it debunks everything you argue?!!
Edit,
ReplyDelete"When a migration from one region to another occurs not every mtDNA lineage takes part in the migration. FOr example, not every mtDNA lineage that existed in the Neolithic Near East moved into Europe. Only some did."
Hence, not every lineage that existed in Africa 60,000 years ago moved into Asia. Only two did: M and N.
@Sam Adams
ReplyDelete"it's the other way around. ALl of us who dis agree with follow science."
As I explained above, geneticists produce data just like artists produce art. And just like you need an art historian to make sense of art, you need an anthropologist to make sense of human origins. That's my turf.
"Most of its subclades are found in Africa not Eurasia. All of L3's relatives: L4, L2, L1, etc, are found in Africa NOT Eurasia. All of that strongly indicates L3, the mother of Eurasian mtDNA, orignated in Africa."
You have it backwards and you got it circularly. L4, L2, L1, are African-specific, they weren't carried out of Africa, hence no out-of-Africa happened. Basal M and N are not found in Africa, they are found on the other side of the world from Africa, hence they didn't originate in Africa. On the other hand, derived M and N lineages are found in Africa.
"The current phylogenetic tree is correct. "
Just like I predicted: you would say something like "the Bible is correct". mtDNA and Y-DNA trees are likely wrong (for the reasons above) because some smart people used wrong assumptions to build them. And then there people like you who consumed them without thinking.
"Eurasian mtDNA is a subset of AFrican mtDNA diversity."
This is from the 1990s. Do you wear bellbottoms per chance, Sam? back-migrations from America to Asia and from Eurasia to Africa are agreed upon by everybody. Hence, no subsets anymore. I challenge people to think about back-migrations as THE migrations.
@ Akraim
ReplyDeleteThat is the whole point. If you look at the dstats I posted, the published pseudo-haploid (less accurate), which is also the one David used in Beagle (because the more accurate diploids were not published. I processed those) share more drift with NW Europeans than the diploid ones.
The diploid ones I processed (the more accurate ones which are closer to the actual Srubna genome) share more drift with E Europeans and Tatars. This makes them a little more "Asian' and a little less "European"
Because we are diploid organisms, the diploid Srubna (8X) is closer to the real Srubna (more accurate) than the published pseudo-haploid one (more artificial and less accurate).
Once I release them in a couple of weeks, and if David wishes to do IBD runs with each of them comparing to various moderns, I can guarantee you a couple of things:
1- There will be considerably more shared haplotypes with moderns, and also longer shared segments (mind you they are not be imputed)than pseudo-haplotypes
2- They should share longer and a higher quantity of haplotypes with E Europeans such as Ukranians than with British for example
Thus if we represent west to east geography by the following line:
Britain ..............................Ukraine..............Recovered Srubna
the diploid Srubnaya I0232 that I processed clusters closer to Ukraninans than to British. It also clusters closer to Tatars than the published pseudo-haploid one does to Tatars.
3- The ordering in the IBD table you referenced should change a little.
@epoch:According to Fu et al the following samples can be sampled as Villabruna + Goyet: La Brana, Loschbour and Ranchot88, while Bichon and Rochedane could be modeled as pure Villabruna. That doesn't match with VNE/VWE distinction you pasted
ReplyDeleteYeah, they say that they need Rank 2 (3 populations) in the supplement *because* its driven by the affinity to GoyetQ116-1 and recent ENA (which they had then) being distinct to some degree (if both patterns were explained by GoyetQ116-1 relatedness, they'd only need Rank 2). Though this does not mean that in part GoyetQ116-1 related ancestry does not affect relatedness to recent ENA people, but it cannot be the only factor...
@Kurd, re your discussion re: IBD with ancient, though would say that at the moment, in Davidski's runs (pseudo-diploid), it's not so much that the Steppe_MLBA samples seem to show raised IBD CM with NW Europe.
Seems more like if you look at the look at the rank ordering, rather they show raised IBD CM with Tajiks and Tatar, however, there is overlap with NW and NE Europe: https://imgur.com/a/GNHUQ
It would make geographical sense if the real diploids matched more closely though, so e.g. Loschbour / Rathlin / Anglo-Saxons would share more drift with W Europe than under p-haploid, Andronovo with Tatars...
@Sam Andrews
ReplyDelete"There's a 0% chance Africa's mtDNA gene pool looked like it does today 60,000 years ago. Therefore the mtDNA impact of the forced migration of Africans to America in recent times is not comparable to the mtDNA impact of the migration of Africans to Asia 60,000 years ago."
That's another pearl. I love your use of math - "0%". Modern African lineages began separating from the "root" of the mtDNA tree 140,000-200,000 years ago. People presumably left Africa 60,000 years ago. It means twice more time elapsed between the onset of the generation of currently observed African genetic variation and the presumed exit from Africa than between the presumed exit from Africa and the present day. We should have had plenty of currently observed African diversity to spread all over the world 60,000 years ago. We don't have anything even remotely similar to this prediction empirically. That's why people came up with a bottleneck idea. Why don't we postulate that only 2 most recently formed lineages left Africa 60,000 years ago. Bingo! This is a special scenario that requires special proof but that special proof has never been presented. Again, as we can see from the post-1492 out-of-Africa, a continent gets sampled pretty deeply and accurately when a migration happens. Nobody just takes two most recently formed lineages with them and just goes to town.
You don't even understand how much of what you think is "science" is just an art of data storytelling.
@Kurd & Matt
ReplyDeleteArkaim is referring to a run that I did with diploid genomes. The one described here...
http://eurogenes.blogspot.com.au/2017/08/a-bronze-age-dominion-from-atlantic-to.html
It shows that the Yamnaya_Kalmykia individuals share somewhat more IBD with Ireland_EBA than with Sintashta, hence I suppose his comment that geography =/= genetics. But the results make sense considering that both Ireland_EBA and Yamnaya_Kalmykia belong to M269, while Sintashta belongs to M417. I expect that Sintashta will share more IBD with M417 samples from the Eneolithic/Early Bronze Age North Pontic steppe.
That run didn't include any Srubnaya individuals. The run that I did with pseudo-haploid samples did include a Srubnaya individual, and it did in fact show that Srubnaya shared more IBD with Ukrainians than with Northwest Europeans. See comments here...
http://eurogenes.blogspot.com.au/2016/01/spatio-temporal-segment-sharing.html
So I'd say everything seems to be in order.
@Matt
ReplyDeleteI thought that they may have been pseudo-haploids, then I realized that they are imputed Martiniano diploids, which BTW are still much much better for IBD work than the published pseudo-haploid versions.
To get an idea of how bad pseudo-haploids are for IBD work, take your own personal genotype data and separate the 2 DNA strands, and replicate 1 strand to create a pseudo-haploid. Take your own pseudo-haploid genome and upload it to Gedmatch, and do a 1 to 1 IBD comparison with your parent. You most likely will get 0 matching segments with your own parent!
With regards to how Martiniano processed his aDNA, I believe he mentioned to me in a discussion a couple of months ago that he removes CT and GA sites to counter deamination. If that is the case then the problem will be a bias and lower hetrozygosity for his set. I didn't have to do that with the diploids I processed.
@German,
ReplyDelete"You have it backwards and you got it circularly. L4, L2, L1, are African-specific, they weren't carried out of Africa, hence no out-of-Africa happened. "
"That's why people came up with a bottleneck idea."
Bottle necks are the norm for mtDNA and Y DNA. Specifically, bottlenecks in which one mtDNA/Y DNA haplogroup becomes popular and dozens of others go extinct, is the norm.
With that in mind, it isn't a big surprise that only one of the many mHGs which existed in Africa 60ky entered Eurasia. It's also not a surprise only some of the mHGs which existed in Eurasia 20ky entered America!!! Yes, I went there.
"Basal M and N are not found in Africa, they are found on the other side of the world from Africa, hence they didn't originate in Africa."
Like how mHG C1b isn't found in Asia and therefore orignated in America. Ok, sure, but mHG C itself definitly orignated in Asia and its distant mother, M, also originated in Asia. Or like how mHG U5 orignated in Europe but its distant mother L3 orignated in Africa.
You got to look beyond derived subclades. To deciver where M/N's ancestors lived don't look at where M and N are found look at where their distant relatives are found: L3a, L3b, L2, L4, L5, L0. They're all found in Africa!!
"mtDNA and Y-DNA trees are likely wrong (for the reasons above) because some smart people used wrong assumptions to build them. "
"wrong assumptions", like what, that recurrent mutationas are extraordinary rare and therefore there's a 0% chance the current phylogenetic tree is wrong? That "assumption" has been proven time and time again.
You arguing my "assumptions" about mtDNA are wrong is like someone arguing pilot's "assumptions" about the laws of physics are wrong.
Just as pilot's put faith into the laws of physics and witnesses their truth everyday I put my faith into assumptions about mtDNA and witnesses their truth everyday.
That's why I'm so shocked by your idiotic claims. I literally see your theories get proved wrong every day in my work.
"This is from the 1990s."
Like how the theory of gravity is from the 17th century. Every academic besides the ignorant or stubborn or ethnocentric (those Chinse guys) still agree with it.
I have Tianyuan. Converting to eigenstrat, then I'll see if anything was missed.
ReplyDelete@Sam Adams
ReplyDelete"Bottle necks are the norm for mtDNA and Y DNA."
Where do you get your norms from? We just saw that one firmly established out-of-Africa migration sampled all of the deep lineages (including A00) on both Y-DNA and mtDNA sides. The first mtDNA studies were conducted using African American samples to represent Africa, remember. So this is a baseline. Under normal circumstances, ancient populations move around without bottlenecks because they are small and substructured. Now, you have to prove that the bottleneck happened because it's a special case, which requires special circumstances to happen. aDNA could resolve this, and it has failed to turn up any African lineages that presumably went extinct. So out-of-Africa was tested and it failed.
"that recurrent mutationas are extraordinary rare and therefore there's a 0% chance the current phylogenetic tree is wrong? That "assumption" has been proven time and time again."
Where do you get your "proofs" I wonder? Yes, if you take Denisovan and Neandertal mtDNAs you can see that for the most part (but not always) they turn up the same character state at the sites that form the "upstream clades" in modern humans as modern Africans. But then we didn't descend from either Denisovans or Neanderthals, so it doesn't really matter. What we need is aDNA evidence from more and more ancient genetic ancestors of modern human populations to establish what the character states were on those lineages. For now, it's quite clear that the most ancient mtDNAs belong to hg R (and now there is an M in Goyet). We just have to go with what's available. There are tons of recurrent mutations on mtDNA. Just look at the mess that Phylotree has created.
Considering that there are no African lineages in ancient Eurasian samples, no evidence in modern Eurasian populations for a migration out of Africa of lineages that are actually attested in modern African populations at the right genealogical depth, a right assumption would be to treat mtDNA L lineages as a subset of mhg M (as Shi Huang does) or N, with mutations going from the character states attested in M and N lineages to character states attested in L lineages. It's just the right logic to build a theory. Why wouldn't they be homoplastic with some states observed in Neandertals and Denisovans? And then keep testing it with more aDNA samples. But out-of-Africa is not even materially represented anywhere in the data to become a theory in the first place.
"Like how the theory of gravity is from the 17th century."
But the subsets were disproven by data in the 2000s. That's the point. Don't try to drag Newton to put a fig leaf on your ignorance.
@Sam Adams
ReplyDeleteAnd don't forget that all of the mutations shared between mtDNA L1 and L3 are not "African" (as out-of-Africanists tried to make everybody believe). They are pan-human and they set us apart from Neandertal and Denisovan sequences. They are found on all lineages and on all continents. They could have arisen anywhere, including America. But what's important about African mtDNAs that they have fewer of those pan-human sites. This suggests that they are derived from non-Africans.
That's where the notion of a subset of relevant. Many African mtdNAs have only a subset of pan-human mtDNA, and that how African lineages have formed. L3 broke off first because they have all the pan-human sites, then in cascading fashion all others. That's how we avoid the nonsensical conclusions that San and Pygmys are million years old and their lineages were isolated from the rest of humans for longer than humans have been a species.
But geneticists are not logicians, they are lab workers with good tools and a lot of money. Almost like elfs.
"between mtDNA L0 and L3" above.
ReplyDelete@German
ReplyDeleteYou're out with the fairies in la la land.
I'm doing you a favor by telling you this, because you're getting your ass handed to you by some kid online called Samuel, and totally humiliating yourself in public. You need to stop this right now.
@German
ReplyDeleteYou know that mtDNA does not undergo recombination, right? So how exactly could the Africans have a sub-set of the pan-human mutations that set us apart from Neanderthals and Denisovans? You think that some of these sites specifically back mutated to the Neanderthal and Denisovan alleles after they seperated from the rest of the humans? Your argument makes zero sense. I would to hear you explain how L0-L2 arose from L3.
@Sam Adams
ReplyDeleteNow back to Amerindians. Their mtDNAs are contrastive to African mtDNAs: Africans have the fewest number of pan-human alleles and the largest number of derived alleles. Amerindians have all the pan-human alleles, all the extra-African alleles but the fewest number of derived alleles in the world. This makes them the most underived of all human populations, hence the most basal. The fact that Amerindians have all the extra-African alleles is further support that they are not just the least derived but also that the split into Eurasian lineages likely started in America. That's why we see MA-1, the basal lineage for West Eurasians and Tianyuan (the basal lineage for east Asians) both show strong affinities to Amerindians. But since East Asians and West Eurasians didn't mix since splitting into two distinct populations, neither of them shows as strong of an affinity to each other's most basal aDNAs as Amerindians.
@Davidski
ReplyDeleteI told you that I don't care about your onions outside of Indo European origins. And the only reason I care about your opinions about Indo-European is because I had expressed them myself long before you. What are you talking about "Samuel hands me my ass"? I'm teaching a kid the right way of thinking. Do you walk backwards, too?
@Karl
ReplyDelete"So how exactly could the Africans have a sub-set of the pan-human mutations that set us apart from Neanderthals and Denisovans?"
Oh, long time no see! Just go to the tree and see for yourself. There are alleles that are found on all continents. Let's call them "L3". It's a pan-human set. They are found in America, Asia, Europe and Africa. Then L6,L5,L4,L2,L1, and L0 have progressively fewer and fewer alleles from that set. That's what "nested subsets" mean. A subset of Africans is characterized by nested subsets of pan-human alleles in which a lot of derived diversity exists.
@ David
ReplyDeleteThanks for the clarification and keep up the good work. Pseudo-haploids are out for IBD work because accuracy at both alleles at a position is very important. ADMIXTURE and ADMIXTOOLS are more forgiving because accuracy at one of the 2 alleles is usually sufficient. Thus I recommend you do a couple of runs using diploid aDNA with moderns.
With regards to imputed genomes, the advertised high imputation accuracies are only applicable to common variants and not rare variants which define populations. Here the accuracy drops off sharply.
I have about 6 steppe samples (unimputed) that I diploid processed. Let me know if you want to give them a try.
Africans have a lot of diversity, but not at your sub-set of so-called pan-human sites that seperate them from L3. At these sites they largely have the same alleles as Neanderthals and Denisovans. This is how you can tell what the actual ancestral allele was. The L3 alleles are derived, not basal.
ReplyDelete@Karl
ReplyDelete" At these sites they largely have the same alleles as Neanderthals and Denisovans. This is how you can tell what the actual ancestral allele was."
But we are humans, a different species, from Neandertals or Denisovans, why would we use them to determine the ancestral states on human sequences? We didn't descend from either of them, we have our own line of descent that we haven't found genetic ancestors of yet. We link back to the ancestors of Neandertals and Denisovans. San and Pygmies are by far closer to the rest of humans than they are to Neandertals or Denisovans, hence if they differ from humans in those pan-human alleles, they just "flipped the switch" there, not retained some kind of plesiomorphies from the time before humans, Neandertals and Denisovans had split from each other.
That's the correct logic and the data is fully consistent with it. The recent skin color paper shows (finally) that San share derived alleles with Europeans and other Sub-Saharan Africans share derived (sic!) alleles with South Asians and Australo-Melanesians.
@German:
ReplyDeleteNot this again! When I first heard your idea of a possible reversed phylogeny I took it seriously and attempted to construct such a tree starting a few key places you mentioned (where Neanderthals/Denisovans share a mutation with non-Africans that Africans lack). As I reported to you back then, it very quickly becomes untenable, requiring nearly an order of magnitude more "duplicate" mutations than the currently accepted phylogeny does.
It's certainly an interesting idea, and hats off to you for creative thinking, but it fails at the very first hurdle... as I've said before you don't need to take my word for it, you can prove it to yourself with about 30 minutes of independent research, literally. I'm a bit surprised you still haven't done that exercise all these years later and are still barking up the wrong tree, as you seem like an intelligent enough person to do it.
I think I may have something not in the paper, but it will take a bit more looking and some graphs.
ReplyDelete@Tobus
ReplyDeleteI'm past that version of tree reversal. I don't think it's methodologically correct to try and root an mtDNA tree in any matches between humans, on the one hand, and Neandertals and Denisovans, on the other. It's just too binary. If you have a mutation once then the only other option is most of the time to go back to the "ancestral" state. It becomes a chicken and egg or a circular argument. Once you have it mutated once (and we have to assume this for pan-human alleles under any scenario), you have to guarantee that it will never switch back again and this is impossible to guarantee, hence you are always in danger of falsifying your own premise with the passing of time. But you can root the tree in the alleles that are pan-human, as defined above, and then, with more progressive aDNA samples keep looking for human direct genetic ancestors and course correct there if needed. The whole phylogeography of human mtDNA (ancient and modern) suggests that Africa can't be a source continent. So the context favors the progressive reduction in the number of pan-human alleles in L6,L5,L4,L2,L1,L0 corresponding to the colonization of Africa from north to south.
@German Dziebel
ReplyDeleteCould you explain this to me?
https://genetiker.wordpress.com/y-snp-calls-for-ust-ishim/
https://genetiker.wordpress.com/mt-snp-calls-for-ust-ishim/
@Arkaim
ReplyDeleteWhat's not clear? Positive calls are in bold, negative are all the rest. If you keep scrolling and scrolling you'll begin finding blocks of bolded alleles - they represent the defining alleles for superclades (e.g., mtDNA L3, N, R) but no subclade alleles, hence he assigned UI to megahaplorgoup (or superclade) R. Randomly occurring bolded alleles are presumably homoplasies.
https://genetiker.wordpress.com/mt-snp-calls-for-ust-ishim/
ReplyDeleteGreat! I didn't know that Genetiker did a calculator also for the mt. Now I'll be able to understand the matter also without computers, programs, etc. You that I work with paper and pencil...
You know... as you see my computer is slower than mi mind!
ReplyDelete@German
ReplyDelete" I don't care about your onions outside of Indo European origins"
So why are you wasting your time on a blog arguing with people who don't agree with you? If you're so convinced that you're right, why not write a paper on it?
It’ wonderful it’s wonderful it’s wonderful
ReplyDeleteA great work of Genetiker. It cannot be done for the Y, because the SNPs are many, not all known, and not all can be tested in the aDNA, but the mt, with its only 16569 markers, is the perfect proof.
Russia Ust'-Ishim, western Siberia M 45530-40610 calBCE 2137615 K2a M526. Initially reported as K(xLT), renamed K2; Derived M2308 (GRCh37 Y: 7,690,182 bp) found by Posnik, shared with NO: new clade created - K2a (Poznik supp. fig. 15.) R* 73G, 263G, 750G, 1438G, 2706G, 3107d, 4769G, 7028T, 8860G, 11719A, 14766T, 15326G, 16150 Fu 2014; Fu 2016; Posnik 2016; Lazaridis 2017
The line of Ust’ Ishim is also mine of course
K1a1b1e-G9932A
born beyond any doubt inTuscany!
My mutation G9932A happened 11 times as a mutation defining a new subclade: 4 from L0d2b to L3f3 and 7 from C1c7 to K1a1b1e, thus also counting L3 as an African line (what at this point isn’t true anymore) the not African lines win statistically 7 against 4. The proof.
ReplyDeleteAnd don't say that not African lines have more people, because African lines should get more time.
@ Bob Floy
ReplyDelete“@German
‘ I don't care about your o[pi]nions outside of Indo European origins’
So why are you wasting your time on a blog arguing with people who don't agree with you? If you're so convinced that you're right, why not write a paper on it?”
The Genius of Kinship (2007) analyzes a database of 2500 kin terminologies to arrive at a number of diachronic universals suggestive of the origin of behaviorally modern humans in the New World
Read this paper of German Dziebel, and next century contact me. We put in the sack pretty all the PhDs of the presumed best universities (I fought for the Italian Refugium and we’ll see next), who is Bob Floy?
Of course it is not a "paper" but a book of many hundreds of pages!
ReplyDelete@ Sam Andrews
ReplyDelete"You're of the same kind as Rob, every Liberal Night Show Host, and half the moderators at Anthrogencia"
Lol. Yeah Falon and I go way back
It's great how you've lumped a whole bunch of people together, kinda like those 'Liberal regressives' you go on about so often, but I guess the irony is lost on you, coz you're a dumbass. Whoops , was that ''snobby' of me ?
Yeah, you got banned from AG for continually, despite warnings, making racist remarks like 'Mongey'.
And you still can't spell, let alone understand the nuances of my points on European history; because you probably couldn't even point to Europe in a map. But you don't see that as a problem because the "liberals" that educated you told you that youre still special, despite your obvious LDs. Again, ironic
So why don't you just stick to counting mtDNAs, and stop your making blanket assertions, like the regressives that you chastise.
And don't use my name in vain, I know you have a crush on me, but I'm not interested.
Mutation T12705C happened once at the mHG R* level.
ReplyDeleteMutation C12705A happened only once in mHG D4o.
Back mutation C12705T actually is the ancestral position and is found 14802 times only in upstream hgs of R*.
At the mHG N* level we have
G8701A, C9540T, G10398A, C10873T, A15301G!
G8701A happened 5 times, but only on N* at the beginning of a main subclade (the others: L0a2a2a1, L1c2a2, L3e1c, D4g1).
C9540T happened only once at the mHG N* level.
G10398A happend 5 times and only in mHG N* at the beginning of a main subclade. The others are: L0b1d1, L1c1a1, L3e1a3, J1c8. J1c8 is signed with a double asterisk, but only for the prejudice that N* is derived and not ancestral.
C10873T happened only once in mHG N*.
A15301G happened only once in mHG N*.
G15301A is documented 5 times: Lc1d1, L2’3’4’6, B4c1b2a2!!, U5a2b4!!, K1b1bE!!. Of course the double asterisk is due to the prejudice that L3* is derived and not ancestral to all the hgs.
K1b2b!!
ReplyDeleteGioiello, L3 is not made derived because people follow the Out of Africa model; people follow the Out of Africa model because L3 is derived. The methods of phylogenetics are used to place the root of the tree, not prejudice. Look them up.
ReplyDelete@Chad: Still have any graphs with Tianyuan to post? What do you think of those in the paper?
ReplyDelete@Kurd: With regards to imputed genomes, the advertised high imputation accuracies are only applicable to common variants and not rare variants which define populations. Here the accuracy drops off sharply.
How much do you reckon this actually mattes? Using linkage even only on common variants should offer quite a big increase over using unlinked common variants...? How much more resolution would you get from rare variants, and is there the data quality in the ancients (currently or feasibly ever) to get it...?
...
Generally, really interested in anyone's thoughts about the paper. Not so interested in "How German Dziebel doesn't understand basic concepts and believes that he has discovered ideas which are unknown to thousands of researchers who are actually specialists in the field (rather than simply being wrong and being unable to publish any of this outside of messageboards and blogs)". Again.
For those who have already forgotten.
ReplyDeletehttp://www.pnas.org/content/110/6/2223
I'm afraid, Gioello, that your own logic hardly rises to that of a turkey. Oh well, enough catering to your senility.
ReplyDelete@ capra internetensis
ReplyDelete“I'm afraid, Gioello, that your own logic hardly rises to that of a turkey. Oh well, enough catering to your senility”.
We’ll see next who will be the turkey at the Middle Eastern banquet…
@Gioiello
ReplyDelete"The Genius of Kinship (2007) analyzes a database of 2500 kin terminologies to arrive at a number of diachronic universals suggestive of the origin of behaviorally modern humans in the New World"
Oh, so he wrote a book? That's awesome.
"who is Bob Floy?"
You mean, like, in a cosmic/ontological sense?
Off topic, but in case anyone is interested, a new paper on biorxiv relating to ancient North American Native dna (and one East Siberian from approx 4000 YBP) - https://www.biorxiv.org/content/early/2017/10/13/203018
ReplyDelete(Not hugely interested in the details of peopling of into America, but perhaps some folk on here are...)
@Matt
ReplyDeleteYesterday, these came out as well.
https://www.biorxiv.org/content/early/2017/10/13/200139
https://www.biorxiv.org/content/early/2017/10/13/200113
Waiting for the South Asian and Bedouin papers.
@Matt Thank you very much!
ReplyDelete@Matt
ReplyDeleteI’m currently investigating this using various case-controls (diploid/haploid/imputed diploid - SGDP data E European/SC Asian) using various MAF thresholds. Incidentally, MAF varies by dataset. For ex if rs1234 has a variant A with MAF of 5% using African set, it can be 20% when using a European set. Also forgot to mention that big gaps in data reduce imputation accuracy.
With regards to LD, I have only used it to remove linked markers in ADMIXTURE to reduce the impact of recent admixture events which tend to increase LD. Haven’t played around with the reverse.
Yes, additional information can be gained by diploid genotyping 3X+ aDNA, in fact it is crucial in IBD work. This has to be done properly, and one has to 1st address post-mortem damage and human contamination which I have done
I will be sharing some of my work regarding the above using heatmaps, graphs, and charts when I publish my article
@Matt
ReplyDeleteThanks, hadn't seen that study. So much at once, I am not even done with the African papers.
Too bad there are no artifacts with Tianyuan Man. Zhoukoudian is nearby, mostly Peking Man stuff; there is small flake tool UP beginning around 34 ka but that is exceedingly generic. Well, maybe if someone here is versed in Chinese archaeology they can tell us something.
There is a contemporary Initial Upper Palaeolithic (blades made on flat-faced cores) in Ningxia, likely related to that in the Altai, but that is 1000 km away, though some vaguely similar stuff is supposed to reach to Korea. If Ust' Ishim represents the North Asian IUP there's no noticeable gene flow to Tianyuan.
I think we are missing something by not modelling modern East Asians as a mixture, but I have no idea how to fix it.
Could calculate the D-stats?
ReplyDeleteMbuti Tianyuan Sunghir1 GoyetQ116-1
Mbuti Tianyuan Sunghir2 GoyetQ116-1
Mbuti Tianyuan Sunghir3 GoyetQ116-1
Mbuti Tianyuan Sunghir4 GoyetQ116-1
Mbuti Tianyuan Sunghir1 Vestonice
Mbuti Tianyuan Sunghir2 Vestonice
Mbuti Tianyuan Sunghir3 Vestonice
Mbuti Tianyuan Sunghir4 Vestonice
Mbuti Tianyuan Sunghir1 ElMiron
Mbuti Tianyuan Sunghir2 ElMiron
Mbuti Tianyuan Sunghir3 ElMiron
Mbuti Tianyuan Sunghir4 ElMiron
Mbuti Tianyuan Sunghir1 Kostenki14
Mbuti Tianyuan Sunghir2 Kostenki14
Mbuti Tianyuan Sunghir3 Kostenki14
Mbuti Tianyuan Sunghir4 Kostenki14
Mbuti Tianyuan Sunghir1 Bichon
Mbuti Tianyuan Sunghir2 Bichon
Mbuti Tianyuan Sunghir3 Bichon
Mbuti Tianyuan Sunghir4 Bichon
Mbuti Tianyuan Sunghir1 Villabruna
Mbuti Tianyuan Sunghir2 Villabruna
Mbuti Tianyuan Sunghir3 Villabruna
Mbuti Tianyuan Sunghir4 Villabruna
Mbuti Tianyuan Sunghir1 Loschbour
Mbuti Tianyuan Sunghir2 Loschbour
Mbuti Tianyuan Sunghir3 Loschbour
Mbuti Tianyuan Sunghir4 Loschbour
Mbuti Tianyuan Sunghir1 Karelia_HG
Mbuti Tianyuan Sunghir2 Karelia_HG
Mbuti Tianyuan Sunghir3 Karelia_HG
Mbuti Tianyuan Sunghir4 Karelia_HG
Mbuti Tianyuan Sunghir1 MA1
Mbuti Tianyuan Sunghir2 MA1
Mbuti Tianyuan Sunghir3 MA1
Mbuti Tianyuan Sunghir4 MA1
Mbuti Tianyuan Sunghir1 AfontovaGora3
Mbuti Tianyuan Sunghir2 AfontovaGora3
Mbuti Tianyuan Sunghir3 AfontovaGora3
Mbuti Tianyuan Sunghir4 AfontovaGora3
Mbuti Tianyuan Sunghir1 Ust_Ishim
Mbuti Tianyuan Sunghir2 Ust_Ishim
Mbuti Tianyuan Sunghir3 Ust_Ishim
Mbuti Tianyuan Sunghir4 Ust_Ishim
Mbuti Tianyuan Sunghir1 Karitiana
Mbuti Tianyuan Sunghir2 Karitiana
Mbuti Tianyuan Sunghir3 Karitiana
Mbuti Tianyuan Sunghir4 Karitiana
@German:
ReplyDeleteI'm past that version of tree reversal
Glad to hear it, it was wrong. :)
I don't think it's methodologically correct to try and root an mtDNA tree in any matches between humans, on the one hand, and Neandertals and Denisovans, on the other. It's just too binary
OTOH, it makes perfect sense. If you're looking for "pan-human" alleles, then those that all humans have, and that neither chimps, Neanderthals or Denisovans have, are exactly what you want.
If you have a mutation once then the only other option is most of the time to go back to the "ancestral" state. It becomes a chicken and egg or a circular argument.
I think this is a mistake - it becomes a *probability* argument not a chicken/egg argument.
While there is a chance a human-only mutation will mutate back to the ancestral state it is very small - even if we assume a 10% chance (which it way too high), each back-mutation required makes the model an order of magnitude less likely. So if defining an allele as "ancestral" requires 10x more back mutations than calling it "derived", you can be sure it's "derived" with 3-sigma confidence.
So the context favors the progressive reduction in the number of pan-human alleles in L6,L5,L4,L2,L1,L0 corresponding to the colonization of Africa from north to south.
I think you are confusing "pan-human" with "pan-continental"... just because an allele can be found on a continent doesn't mean that all humans in that continent have it. True "pan-human" alleles are shared by all moderns, and absent from non-humans - hint, if it defines a modern haplogroup, it's not pan-human.
If you try to root a tree using modern variations you are doomed to fail - by definition no modern variation can be the root.
@supernord
ReplyDeletehttps://drive.google.com/file/d/0B8XSV9HEoqpFemxxZ0F0b1JlY2M/view?usp=sharing
@Tobus
ReplyDeleteIt's not a probability problem. It's an absurdity test. There's a clear logical and historical pattern here. Most humans on all continents (America, Australia, Asia, Europe, Africa) share a set of pan-human, pan-continental mutations. If there're modern outliers located on just one continent that show a progressive reduction in the number of those alleles in that very same set, then it stands to reason that the latter are derived from the former through lineage branching. That's what a subset of a set is. There's no way around it.
We can't imply that San and Pygmies have retained plesiomorphies from the time before the split into HSS, Denisovans and Neandertals, without denying the unity of the human species. I assume that modern humans are a species, that they have accrued certain mutations after they split from their MRCA and before they began splitting into continental populations as part of the colonization of the world.
"If you try to root a tree using modern variations you are doomed to fail - by definition no modern variation can be the root."
This is nonsense. A pan-human set of alleles is the only logical way to root alleles found in continental populations. Humans is a species, Africans, Asians, etc. are a subspecies. There's no such thing as San+Pygmies+Denisovans+Neandertal clade.
@Capra
"The methods of phylogenetics are used to place the root of the tree, not prejudice."
What methods of phylogenetics are you talking about that are different from the ones a professional such as myself is espousing?
@German - could you just give some hard data instead of these arguments? Numbers please.
ReplyDelete@ Ryan
ReplyDeleteWhilst I am waiting that German as a “professional” answers you, I make you note that those I gave above were “numbers”. With some approximations (as to numbers of peoples involved and times) I considered some very slow mutating markers demonstrating that they are explained better by putting L3 at the origin of the tree.
@Ryan
ReplyDeleteThere's no dearth of numbers in genetics research. What's missing are good arguments. I've got some. I'm just waiting for them to sink in. Then we can revisit the numbers.
One of my trees with Tianyuan
ReplyDeletehttps://drive.google.com/file/d/0B962TtPkX1YnUVd4WWI5TnhwTXc/view?usp=sharing
@German
ReplyDeleteOk. Now I get it. So are chimps also highly derived from modern humans? Did they just have a ton of back mutations from the pan-human state to the ancestral allele state? If not, how can you tell the difference?
Chimpanzees belong to a different genus and species than AMH. Homo erectus is the first specimen known to have some arcahic but more modern human ancestral allele's.
ReplyDeleteDuh, Karl_K, you're using fast-mutating mutations. Those just mutate at a user-defined speed. You have to use the slow-mutating mutations, you can recognize them because they support the phylogeny you already know is correct.
ReplyDeletePersonally though I think the entire Eurasian tree is under a certain subclade of L2a1. On EgyptSearch I was told that white people are white because of a cosmic ray burst, since radiation is known to cause albinism. Logically that would lead to a string of mitochondrial mutations as well. There is however an alternative theory that white people are white because they spent the entire glacial period living in caves. But perhaps radon levels were high?
@ capra
ReplyDelete“You have to use the slow-mutating mutations, you can recognize them because they support the phylogeny you already know is correct”.
Not only ”ragiona come una capra” but gets also the “spirito di patata” this “capra dal viso gentile”, or best a “coacervo” like “internetensis”.
The slowest mutations markers were used from Shi Huang, and are the unique from discriminating a measurable number of mutations. You may have another proof, when your friends come back from the banquet (if) and publish what they have in their labs.
@German:
ReplyDeleteMost humans on all continents (America, Australia, Asia, Europe, Africa) share a set of pan-human, pan-continental mutations
I agree with this statement, but I suspect the specific mutations you are talking about are *not* the ones that "most humans all on continents .. share". Perhaps you can specify which mutations you are suggesting we should root the phylogeny with?
mHG L0 begins with these mutations: G263A, C1048T, C3516a, T5442C, T6185C, C9042T, A9347G, G19589A, G12007A, A12720G.
ReplyDeletemHG L0 has nothing to do with all the other hgs.as yHG A00 has nothing to do with all the others hgs, and very likely the oldest Africans were mHG L0 and yHG A00. Mutation C3516a is found only in mHG L0, and the unique mutation in classifying a subclade is C3516T in mHG A2a1.
Of course L0 doesn’t descend from L3.
Also A9347G happened only in L0. Thus we have 9357A unchanged from hundreds of thousands years in our line.
ReplyDeleteAnd these two (see above) are synonymous mutations, thus not subjected to selection.
@ Davidski, could you calculate f3(Tianyuan,X;Mbuti) for this list: http://textuploader.com/d4h83?
ReplyDeleteJust for the sake of completeness really so I can compare them to your previously published f3 stats for Ice Age Europe.
Also for D / f4 stats, if you have Tianyuan in the right datasets, would it be possible to run:
A - http://textuploader.com/d4xa7
B - http://textuploader.com/d4xaw
I'm interested in A particularly as though the paper included quite a few stats, there were also few who were not covered and I want to see if there are any evident clines across East Asia (either with African outgroup or when using Denisovan / MA1 as outgroup). A includes some insular SE Asia populations who may have ancestry edges related to Onge.
@ Chad, thanks.
Looking at that tree - and this is nothing against your work on it, because of course it produces the correct stats - but the one bit I do find a little tricky is how the "pOnge"(your graph pO1) ancestry in Onge and Surui is from populations which:
a) share a little more drift with Ami than Tianyuan (e.g. E1a)
b) form the outgroup to which both Tianyuan and Ami are a clade (e.g. Oceania)
Obviously it does work somehow! It's just not a phylogeny where I would look at it and then intuitively go, "Oh, yeah, that explains the (D(Surui/Chane, Mixe, Tianyuan, Mbuti) = 0.02, Z > 3 result" specifically. Just how does ancestry from an outgroup to Tianyuan and Ami and from a sister population of Ami result in Surui having an extra affinity to Tianyuan relative to Mixe, I don't understand...
It also seems like it should imply D(Surui, Mixe, MA-1, Mbuti) and D(Surui, Mixe, MA-1, Mbuti) are positive (ANE ancestry in Surui is diluted relative to Mixe by Onge like ancestry).
Still @Chad's tree, just doing a sense check of how well this does this fit with their supplement Table S2f (-3<>3 is significant):
ReplyDelete1.
D(Tianyuan, Ami; Onge, Mbuti) = Z: -5.9
D(Tianyuan, Ami; Papuan, Mbuti) = Z: -3.2
(Relative to Tianyuan, Ami share more of the Onge and Papuan specific drift paths from Mbuti)
2.
D(Ami, Onge; Tianyuan, Mbuti) = Z: 4.7
D(Ami, Papuan; Tianyuan, Mbuti) = Z: 6.3
(reversed stats would be:
D(Onge, Ami; Tianyuan, Mbuti) = Z: - 4.7
D(Papuan, Ami; Tianyuan, Mbuti) = Z: -6.3)
(Relative to Onge and Papuan, Ami share more of the Tianyuan specific drift paths from Mbuti)
3.
D(Tianyuan, Onge; Ami, Mbuti) = Z: -1.3
D(Tianyuan, Papuan; Ami, Mbuti) = Z: 3.5
(Relative to Tianyuan, Onge does not share more or less Ami specific drift path from Mbuti
Relative to Tianyuan, Papuan shares less of the Ami specific drift path from Mbuti)
I guess:
1: Resolves via admixture from Ami sister into Onge / Papuan.
2: Resolves by having Tianyuan and Ami form a clade relative to the majority ancestry in Onge / Papuan.
3: 3 resolves via a greater edge from the Ami sister into Onge than Papuan?
So it seems all these stats would be consistent with the tree...
@Matt
ReplyDeleteThe stats can only be run with the following modern pops at the moz.
https://drive.google.com/file/d/0B8XSV9HEoqpFNXhUcW1lNlVMX0k/view?usp=sharing
In everyone's comments about the obvious evidence against any "Out of the Americas" or "Out of East Asia" (or Eurasia in general) theory of Anatomically Modern Human origins, everyone here forgot one major piece of evidence: The mtDNA haplotype of Tianyuan himself.
ReplyDeleteDNA analysis of an early modern human from Tianyuan Cave, China, Fu et al. (2012)
Tianyuan is a basal haplotype of mtDNA, at the root of Central and East Asian, Polynesian, and Native American (Amerind) mtDNA B.
Tree of the Tianyuan and 36 present-day mtDNAs belonging to haplogroup B.
There's no ifs, ands, or buts about it:
About 25% of Native American mtDNA (in haplogroup B, under haplogroup N), and many East Asian and Oceanian mtDNA haplotypes derive from Tianyuan - or an extremely close relative - just like Eurasian mtDNA N derives from Oase1's mtDNA pre-N* (derived for 3 out of 5 mtDNA SNPs. Oase1, as we know, is Y-DNA K2a1-M2335*, otherwise known as "NO1", and the same Y-DNA clade as Ust'-ishim. As we also know, Oase1 was the earliest sequenced European Anatomically Modern Human, and in fact 1/8 Neanderthal.
Siberian Ust'-ishim was mtDNA R*, and this gives us the mtDNA sequence of:
1. Oase1 in pre-N*
2. Siberian Ust'-ishim in R*
3. Tianyuan in B*
Therefore, aside from any archaeological, linguistic, autosomal, and Y-DNA evidence, the current early pre-40,000 BP mtDNA evidence is a clear record of the settlement of Eurasia and the Americas, with the East Asia later than Central Asia, and the Americas last.
Any "theory" to the contrary must explain this evidence, including the mtDNA result of Tianyuan, without resorting to extra-scientific ideas such as conspiracy theories.
@OpenGenomes, you are being too kind to Dziebel. While I do not like to comment on political tangents on this blog, the man has stated, in this thread, (in one of the few parts of his stream of verbiage that stood out): But don't start squealing that Shi Huang is an agent of Chinese imperialism. Out-of-Africa is just as much as product of Western liberalism as out-of-Asia is a product of Chinese imperialism..
ReplyDeleteThat is, going well beyond the idea that there are simple biases in our culture towards an African origin of modernity (as is obvious and has been observed by firmly mainstream folk like Milford Wolpoff), to state that the established, mainstream science, is nothing more than a complte reification of ideology and interests. If he really does mean this, this is Lysenkoism.
In its own way I would be unsurprised if this was equally a turnoff to scientists reading this blog as Gioello's detours into barely intelligible anti-semitism, or batman's ersatz Nordic fantasy fiction prehistory.
@ Gioello, you are hardly in a favourable position to accuse others of illiteracy in the English language.
If nothing else, my apologies for writing your name wrongly, Gioiello.
ReplyDelete@Grey, Ed Yong (who appears to have fallen in quality over the last ten years, since his Not Exactly Rocket Science days, if that article is any estimation) got a quote from a moron at the end there. Even if she talks a lot of nonsense, Sarah Tishkoff would not have said that - it's not even clear the study even shows that Europeans have more ancestral variants at various sites, just that these are responsible for the within African variation.
ReplyDeleteFor whatever the article's speculation that these dark variants arose in Erectus, it's also not really clearly evident that these alleles didn't exist in balancing selection with the proto-human Great Ape population - chimpanzees, even within Troglodyes alone, and within every subspecies, frequently have dark unexposed skin. These do not seem so likely to have been an innovation in chimpanzees. It's a bit Just So to assume these variants would all have been removed by selection prior to the early human ancestors.
Don't take it as anything against the scientists that Yong chose to get a quote from a guy who wasn't involved.
@Matt
ReplyDelete"Don't take it as anything against the scientists..."
i think it's a media thing.
@Gioiello
ReplyDeleteI have to ban you now, because you keep breaking the "no conspiracies" rule.
You got plenty of warnings from me, so you don't have anything to complain about. I've added your name to the banned commentators list.
http://eurogenes.blogspot.com.au/2017/09/banned-commentators-list.html
It's over. Please find somewhere else to post.
@Arkaim
ReplyDeleteReminder: no politics on this blog.
Also banned is Mr Snow, a stupid troll with no arguments and nothing even remotely interesting to say.
ReplyDeleteI'll also use this opportunity to give final warnings to Azarov Dmitry and German Dziebel.
If you guys don't immediately start posting some coherent thoughts and strong arguments, backed up by hard data, then don't bother posting here again.
Ooooooh.. I think I found more stuff! It seems that even Sunghir looks to have Tianyuan ancestry. Z right around 3, compared to Kostenki14. GoyetQ likely has Tianyuan.. and MA1 may have Tianyuan, plus from further on down the line from stuff closer to the Ami and what makes up most of Native Americans. The more I dig into this, the wilder it seems to get.
ReplyDelete@Chad
ReplyDeleteCool. On the stats supernord requested Sunghir4 turns up significantly further from Tianyuan than the other Sunghirs - which sure makes it look like the others have Tianyuan. Unless that's some kind of artifact?
@ Davidski
ReplyDeleteIf you have time and will, I'd like to ask for F3 stats. Feel free to delete this comment after reading it.
Tianyuan Abkhasian Mbuti
Tianyuan Basque_French Mbuti
Tianyuan Belarusian Mbuti
Tianyuan Bulgarian Mbuti
Tianyuan Estonian Mbuti
Tianyuan Finnish Mbuti
Tianyuan French Mbuti
Tianyuan Georgian Mbuti
Tianyuan German Mbuti
Tianyuan Greek Mbuti
Tianyuan Hungarian Mbuti
Tianyuan Karelian Mbuti
Tianyuan Latvian Mbuti
Tianyuan Lezgin Mbuti
Tianyuan Lithuanian Mbuti
Tianyuan Norwegian Mbuti
Tianyuan Polish Mbuti
Tianyuan Sardinian Mbuti
Tianyuan Spanish Mbuti
Tianyuan Swedish Mbuti
Tianyuan Ukrainian_East Mbuti
Tianyuan Ukrainian_West Mbuti
GoyetQ116-1 Abkhasian Mbuti
GoyetQ116-1 Basque_French Mbuti
GoyetQ116-1 Belarusian Mbuti
GoyetQ116-1 Bulgarian Mbuti
GoyetQ116-1 Estonian Mbuti
GoyetQ116-1 Finnish Mbuti
GoyetQ116-1 French Mbuti
GoyetQ116-1 Georgian Mbuti
GoyetQ116-1 German Mbuti
GoyetQ116-1 Greek Mbuti
GoyetQ116-1 Hungarian Mbuti
GoyetQ116-1 Karelian Mbuti
GoyetQ116-1 Latvian Mbuti
GoyetQ116-1 Lezgin Mbuti
GoyetQ116-1 Lithuanian Mbuti
GoyetQ116-1 Norwegian Mbuti
GoyetQ116-1 Polish Mbuti
GoyetQ116-1 Sardinian Mbuti
GoyetQ116-1 Spanish Mbuti
GoyetQ116-1 Swedish Mbuti
GoyetQ116-1 Ukrainian_East Mbuti
GoyetQ116-1 Ukrainian_West Mbuti
Kostenki14 Abkhasian Mbuti
Kostenki14 Basque_French Mbuti
Kostenki14 Belarusian Mbuti
Kostenki14 Bulgarian Mbuti
Kostenki14 Estonian Mbuti
Kostenki14 Finnish Mbuti
Kostenki14 French Mbuti
Kostenki14 Georgian Mbuti
Kostenki14 German Mbuti
Kostenki14 Greek Mbuti
Kostenki14 Hungarian Mbuti
Kostenki14 Karelian Mbuti
Kostenki14 Latvian Mbuti
Kostenki14 Lezgin Mbuti
Kostenki14 Lithuanian Mbuti
Kostenki14 Norwegian Mbuti
Kostenki14 Polish Mbuti
Kostenki14 Sardinian Mbuti
Kostenki14 Spanish Mbuti
Kostenki14 Swedish Mbuti
Kostenki14 Ukrainian_East Mbuti
Kostenki14 Ukrainian_West Mbuti
Tianyuan Kostenki14 Mbuti
Tianyuan GoyetQ116-1 Mbuti
GoyetQ116-1 Kostenki14 Mbuti
@Arza
ReplyDeletehttps://drive.google.com/file/d/0B8XSV9HEoqpFOWd4V3VfZmp4ejA/view?usp=sharing
@All
ReplyDeleteIdiotic post by some sock puppet known as Nomic Belief removed.
ReplyDelete@Tobus
". Perhaps you can specify which mutations you are suggesting we should root the phylogeny with?"
This is easy. Go to Phylotree and start looking at clade defining mutations. E.g., L1,2,3,4,5,6 is defined by C146T C182T T4312C T10664C C10915T A11914G G13276A G16230A. Then in addition to those L2'3'4'5'6 has C152T A2758G C2885T G7146A T8468C. Then L2,3,4,6 adds the following C195T A247G A825t T8655C A10688G C10810T G13105A T13506C G15301A A16129G T16187C C16189T. Then G4104A A7521G from L3'4'6. Then T182C! T3594C T7256C T13650C T16278C from L3'4 and A769G A1018G C16311T from L3.
As we go along, more and more humans have those mutations but fewer and fewer Africans. But then it stops and a large number of Africans (L3-carriers) plus all of the non-Africans will still have them. So the idea is that all of these alleles combined (save a few of the recurrents) are "pan-continental and pan-human". Then L0, L1, L2, L4, L5, L6 would carry a subset of those, plus a large number of unique derived mutations.
ReplyDelete@Open Genomes
"There's no ifs, ands, or buts about it:"
That original paper on Tianyuan showed that Tianyuan shares more nucleotides with Karitiana than with any other population in the world.
Amerindians have sampled each and every mtDNA mhg (B from R, CD from M and A, X from N) that's found outside of Africa. But the number of the lineages attested in America for each mhg is very small. So the question is why a) did they sample all of the mhg, but b) didn't sample any of them more to reflect the much greater diversity of those lineages in the Old World; c) sampled only lineages that are rare to non-existent in modern Siberians but are geographically widespread in the Old World (B, D4h3 in SE Asia, X in West Asia and North Africa, A in Tibet, etc.).
So to the problem with the attestation of African lineages outside of Africa, we have a problem of the attestation of extra-African lineages in America.
@Davidski
ReplyDelete"If you guys don't immediately start posting some coherent thoughts and strong arguments, backed up by hard data, then don't bother posting here again."
All of my thoughts are coherent. Yours are mostly not. That's why I have 2 docs and you have 0. You've taken advantage of the fact that anybody could start a blog. But in my book you need to have something bigger than "login credentials" to speak on the topics you're covering here. I don't care if you "ban" me. You've been pre-banned from all rational communications in any areas of life outside of the web where anything goes and nameless people share their ideas. And instead of harassing me about this or that, you better clean up your comments string of all the meaningless posts that most of your readers are writing. Look at my blog and follow suit: clean, focused and without the endless comments. It takes me a while to plow and extract what to respond to from an avalanche of undereducated posts here. Save your and, most importantly, my time. Do your job as a moderator!
Also, you may want to try and apply and get accepted to Harvard, as someone who's quite similar to Dienekes did. :) :)
@Matt
"the established, mainstream science, is nothing more than a complte reification of ideology and interests. If he really does mean this, this is Lysenkoism."
You are being hysterical just because someone is touching the gods you are worshipping. I love how amateurs start online squealing when science is being treated unfairly in their mind. I just don't like when people speak of out-of-China as an ideological construct, while out-of-Africa is presumed to be "pure science." The degree of the severity of the cultural bias in one vs. the other is subject to debate, but we just have to be aware of the fact, as anthropology teaches us, that scientific discourse is subject to cultural biases.
But the very fact that you and people like you get so inflamed by the comparison tells me that out-of-Africa does have a parascientific significance for people, just like making a cartoon about a revered religious figure sets out a flurry of threats in a religious community.
@Matt
ReplyDeleteRe: Lysenkoism. Lev Zhivotovsky who you may have heard of recently published a book in Russian about Lysenko in which he argued that Lysenko was not such a clear cut anti-hero as we're used to think. So, don't use lightly the words that require a lot of background reading in multiple languages to use appropriately. Every time an amateur clings to something in science that will give his troubled soul a solace, scientists open a debate about it.
@German, not gonna read your screed, but you are very much an amateur on this topic. No qualifications in the field, no published work in the field, no job in the field, no recognition in the field, no reputation in the field, no expertise in the field. You're pretty much on the level of Gioiello, and however many blog comment sections you spam up, you always will be.
ReplyDelete@Matt
ReplyDeleteYou don't know what a "field" is. A field is anthropology of which "anthropological genetics" is a subfield. I have credentials in the field, publications in the field, I have a job in the field, etc.
Could you please go to school for a change? As of now, you can't use relevant words properly.
@German Dziebel
ReplyDeleteYou're banned. Go see a shrink instead of posting on the internet.
Fuck off.
@Davidski
ReplyDeleteFarewell! This will save me a lot of time.
@German:
ReplyDeleteThis is easy. Go to Phylotree and start looking at clade defining mutations
It's as I thought, you are suggesting rooting in the "non-pan-human" alleles. Each mutation defined in Phylotree is only found in subset of humans... think about it. All the mutations listed are *within human* variation - none of them can be "pan-human", by definition.
Take a look at Table S2b.1 of Prufer et. al. 2014 (page 12). Human mtDNA has, on average, nearly 1500 mutations that separate us from Chimps, and about 200 mutations that are distinct from Neanderthals. In comparison the largest human-human difference is just under 100 mutations. The human-human differences are what are detailed in Phylotree, but the alleles that are used to root the tree are the human-Neanderthal/human-Chimp differences - the ones that all humans have, that differentiate us from other species... the truly "pan-human" ones.
As we go along, more and more humans have those mutations
Are you sure about that? Doesn't each subsequent mutation split the tree, so less and less humans have each mutation as we go along? Like everybody has all the basal L mutations, but only some people have the L3-specific ones. Only some of these L3 people have the M mutations, and only a small %age of these have the R1a1b1 mutations. I think you have somehow got yourself back to front on this.
Cross topic, you said to OpenGenomes:
That original paper on Tianyuan showed that Tianyuan shares more nucleotides with Karitiana than with any other population in the world.
That original paper only sequenced chromosome 21, about 1.5% of the genome. Now that we have the full genome, you should update your priors.
@Tobus
ReplyDelete"the ones that all humans have, that differentiate us from other species... the truly "pan-human" ones."
You only prove my point. The "truly pan-human ones" are the ones that ALL of humans share vs., say, Neandertals. The next step are "pan-human ones" MOST humans share that some humans (only in Africa) don't share. So, it's a nested set. You just took one step up the tree, which is fine. The current mtDNA tree goes from "truly pan-human ones" to the ones that are found in L0, L1, L2, L4, L5, L6 and only then to the ones that most Africans and all of non-Africans share. It leapfrogs the level of MOST HUMANS. The mistake comes from the confusion in what "basal" is in a phylogeny. People construct trees that have branches but no trunk.
Ultimately it's the same pan-human sites that just got flipped in some humans who live in Africa.
"Doesn't each subsequent mutation split the tree, so less and less humans have each mutation as we go along?"
I just phrased it awkwardly the first time: if you have a sample and you start assigning sequences with those mutations to Africans vs. non-Africans, you will see that more and more non-Africans would have all of those mutations, while fewer and fewer non-Africans from that sample would have all of them.
@Tobus
ReplyDelete"That original paper only sequenced chromosome 21, about 1.5% of the genome. Now that we have the full genome, you should update your priors."
Of course. I was just referring to the paper that OG brought up.
@German
ReplyDeleteI told you already, go and seek some professional help.
@All
ReplyDeleteThis is where Tianyuan clusters in the Global 10 PCA (look for the red plus).
https://drive.google.com/file/d/0B9o3EYTdM8lQaG9EZlloa25kTjQ/view?usp=sharing
The updated Global 10 datasheets with Sunghir and Tianyuan are here.
https://drive.google.com/file/d/0B9o3EYTdM8lQd1hLRFl4OUdiME0/view?usp=sharing
https://drive.google.com/file/d/0B9o3EYTdM8lQb0ZjM0pLVmtvTFE/view?usp=sharing
@Tobus
ReplyDeleteCorrigendum: "while fewer and fewer Africans from that sample would have all of them."
@Tobus
ReplyDeleteGo and finish your discussion with German at his blog. I don't want the comments here cluttered up with his nonsense.
With all these new ancient samples, I think it would be a good idea to revisit the kind of under-discussed Oase1. Davidski, can you run a few d-stats? Thanks in advance =)
ReplyDeleteMbuti Oase1 Ust_Ishim Kostenki14
Mbuti Oase1 Ust_Ishim Vestonice
Mbuti Oase1 Ust_Ishim GoyetQ-116
Mbuti Oase1 Ust_Ishim Sunghir1
Mbuti Oase1 Ust_Ishim Sunghir2
Mbuti Oase1 Ust_Ishim Sunghir3
Mbuti Oase1 Ust_Ishim Sunghir4
Mbuti Oase1 Ust_Ishim Tianyuan
Mbuti Oase1 Ust_Ishim MA1
Mbuti Oase1 Ust_Ishim AfontovaGora3
Mbuti Oase1 Ust_Ishim ElMiron
Mbuti Oase1 Ust_Ishim Villabruna
Mbuti Oase1 Ust_Ishim Bichon
Mbuti Oase1 Ust_Ishim Loschbour
Mbuti Oase1 Ust_Ishim LaBrana1
Mbuti Oase1 Ust_Ishim Karitiana
Mbuti Oase1 Ust_Ishim Han
Mbuti Oase1 Ust_Ishim Onge
@human443
ReplyDeletehttps://drive.google.com/file/d/0B8XSV9HEoqpFcTh3S3FkblJtdTQ/view?usp=sharing
ReplyDelete>Someone really should get Chokhopani/Sherpa and Jomon/Ainu genomes and run them together with Amerinds, Tianyuan, Papuan and Onge to get a clear phylogeny of structure in the presently known branches of ENA.
I recommend also Paniya, Mala, Australian Aborigines, Oase1, Ust'-Ishim, Sunghir, Lapita Vanuatu, and Sanganji/Hokkaido/Tohoku/Kanto Jomon.
Lapita might be somewhat related to that pre-Ami admixture in ANE.
ReplyDelete>I think we are missing something by not modelling modern East Asians as a mixture, but I have no idea how to fix it.
East Asians are certainly admixed. Just much less so than Europeans and Middle Easterners. I think modern East Asians are a neolithic Mix between populations related to Devil's Gate and Lake Baikal hunter gatherers. Many also have additional Southeast Asian admixture containing ASI elements and a little bit of ANE.
You can try using Devil's Gate, Ulchi, Sherpa, Amerindians, Jomon, Lapita, Ami, AG3/AG2, Lake Baikal, Onge, Aeta, She, Paniya, Harappans and Papuan samples to model them. All samples should be available on Gedmatch except perhaps Jomon, Baikal and Harappa.
@Unknown
ReplyDeleteI think Devil's Gate is recent enough that most of the Lake Baikal element should already be accounted for. I am looking more for an ancient southern (Central Chinese?) element. Could be wrong mind you.
ReplyDelete>I think Devil's Gate is recent enough that most of the Lake Baikal element should already be accounted for. I am looking more for an ancient southern (Central Chinese?) element. Could be wrong mind you.
Devil's Gate and Baikal hunter gatheres likely had partially overlapping genotypes but they weren't the same so both should be tested. Anyway, was the study about Baikal hunter gatherers autosomal DNA finally released? I can't find anything other than a paper about their Y-DNA.
If you want a good example of the southern mongoloid then Lapita Vanuatu fits description. They seemed to be pure southeast asians without any(or least with very small) australoid, negrito or east esian admixture.
You can try Chokhopani1, too.
>Despite modern Europeans having "Basal Eurasian" ancestry (and also speculatively, some ancestry from ENA clade), they still seem closer by outgroup f3 to K14 than East Asians today to Tianyuan.
Could that be due to UP European related admixture in prehistoric Middle Easterners who mixed into Europeans? Either way, those prehistoric populations don't have simple tree-like histories of admixture but complex patterns with multidirectional gene flow.
ENA in Europeans isn't speculative. We know ANE/EHG/WHG/SHG/CHG all had it and they mixed into modern Europeans. Many modern Europeans also have additional recent Siberian/East Asian/Amerindian admixture on top of all that.
@Unknown: "East Asians are certainly admixed. Just much less so than Europeans and Middle Easterners."
ReplyDeleteI think this will be true in the sense that the splits between the ENA subgroups who contributed to East Asians will inevitably have deep less time depth that the splits of Basal Eurasians, and subclades from West Eurasian. It's just sort of almost inevitably true due to the direction of Out-of-Africa and early AMH population structure and where populations would have fanned out.
I would guess West Eurasians have ancestry from populations who split around 60-40k out of Africa, then combined in different admixtures all until the present, while East Eurasians only ancestors who split around 50-40k (then combined in different admixtures all until the present)?
There is also the question of less admixed vs less variable in admixture. The Han Chinese, for instance, are not very variable in relationships to outgroups, but this means they are not variable in whatever admixtures would drive relatinships to outgroups, more than that the population shares a single population history, necessarily.
@Unknown: Could that be due to UP European related admixture in prehistoric Middle Easterners who mixed into Europeans? Either way, those prehistoric populations don't have simple tree-like histories of admixture but complex patterns with multidirectional gene flow.
ENA in Europeans isn't speculative. We know ANE/EHG/WHG/SHG/CHG all had it and they mixed into modern Europeans. Many modern Europeans also have additional recent Siberian/East Asian/Amerindian admixture on top of all that.
Yeah, it's something like that, though I don't know about directions of population movements implied. I would guess it means that Europeans share slightly more of a tree like history with Kostenki than present day East Asians do with Tianyuan, despite complexities of tree structure.
ENA is still under consideration, and they are doing things with different models. I think it's honestly the best model at the moment though.
@Open Genomes, this would be pretty cool if so. Do we not already have Neanderthal Y chromosomes though (from Neanderthal)?
@OG
ReplyDeleteI'm going to ask Qiaomei Fu. Better to hear what's going on from the main source, than to speculate too much.
Alright, the SNP capture chip they're using doesn't offer enough data to infer the Y-haplogroup. They're going to try to capture a lot more data from the Y-chromosome, basically deep sequence, and go from there.
ReplyDelete@David, thanks for asking Qiaomei Fu.
ReplyDeleteThat makes sense too. Will they be using a new custom Y enrichment capture array to do this, or just more shotgun sequencing?
The Reich Lab capture array leaves something to be desired, and is just based on what ISOGG lists as SNPs, rather than trying to capture all the unique regions of the Y (e.g. the Poznik (2013) 100 bp paired-end unique regions). A new, more complete custom Y chip would be of general usefulness, particularly for some of the very low coverage aDNA samples the Reich Lab has already published.
Do you know what method they're going to use?
If the Reich Lab is going to go back and try to capture more of the Y-DNA and publish it separately, this would be the very first time that they did so.
ReplyDeleteThey already use a custom mtDNA capture array, and while the 1240k chip does have Y SNPs, at times it just doesn't capture enough from the Y to include or rule out key subclades.
I guess a 40 kya individual was enough motivation for them to finally create a specific capture array.
Still, you'd think they would have found something general regarding the Y haplogroup, unless what they have is in fact so unusual that every SNP is negative, i.e. it's an Archaic Y.
@David, can you ask Qiaomei Fu if the Reich Lab intends to revisit some of their other samples with a new, much more comprehensive Y-specific chip? That would be incredibly useful, and I think they may be beginning to understand the usefulness of the Y haplogroups and tMRCAs for their analysis, like we've seen now with Sunghir, who are all in an extinct branch of C1a2-V20, as is La Brana1.
The question of course is whether these populations survived directly at all, or not.
At least with some of the Yamnaya samples, with Eske Villerslev's WGS, we were able to determine that there is a tiny R-KMS75 (R-Y20993) clade that appears to be the direct descendants of the late Yamnaya kurgan individuals:
YFull R-Y20993 (KMS75)
The random 2013 ISOGG tree based 1240k chip just isn't isn't adequate any longer, compared to what the other labs are doing. If the Reich Lab develops a new Y-specific capture chip, then we might have some aDNA Y studies on the horizon.
Anyway, regardless of whether Tianyuan has an Archaic Y or not, I think they'd want such a Y capture array to try to sequence a Neanderthal Y, too.
@OG
ReplyDeleteI don't know what they'll be using exactly. But whatever it is, it will be more powerful than the current aDNA chip. We just have to wait for the next paper.
@Matt ”I would guess West Eurasians have ancestry from populations who split around 60-40k out of Africa, then combined in different admixtures all until the present, while East Eurasians only ancestors who split around 50-40k”
ReplyDeleteThere are at least two potentially important factors that affect the estimations of the level of differentiation in ENA populations: pincer model of the colonization of East Asia and the early dispersal of AMH out of Africa to Southeast Asia and Sahul c. 80 000 - 60 000 years ago.
Of course, ENA populations have deep less time splits if the colonization happened only via one southern colonization c. 60 000 years ago. However Tianyuan seems not support this simple model if 1) Tianyuan is divergent from the groups that led up to present day East Asians; 2) Tianyuan is more related to Suruii than to Mixe; 3) Tianyuan prefers East Asian over Onge, Onge prefers East Asian over Tianyuan and East Asian has no preference between Tianyuan and Onge.
HLA loci support the pincer model in which East Asians are the result of a colonization from the north and from the south (http://media.springernature.com/full/springer-static/image/art%3A10.1186%2Fs12862-015-0512-0/MediaObjects/12862_2015_512_Fig1_HTML.gif).
”Our results show that the Southern-origin model can be rejected as it is almost never able to reproduce the observed data in East Asia, whatever the initial conditions are. A formal comparison of the three models using the ABC approach indicates that both two-route models (Pincer and Overlapping) are significantly more likely than the Southern-origin model (>96 % among 1,500 retained simulations) for the three HLA loci.” (https://bmcevolbiol.biomedcentral.com/articles/10.1186/s12862-015-0512-0)
”HLA alleles appear to be unevenly distributed: some alleles observed in NEA populations are widespread at the global level, while some alleles observed in SEA populations are virtually unique in Asia. The HLA genetic variation in East Asia is also characterized by a decrease of diversity from north to south, although a reverse pattern appears when one only focuses on alleles restricted to Asia. These results reflect a more complex migration history than that illustrated by the "southern-origin" hypothesis, as genetic contribution of ancient human migrations through a northern route has probably been quite substantial. We thus suggest a new overlapping model where northward and southward opposite migrations occurring at different periods overlapped.”
(https://www.researchgate.net/publication/51043489_Challenging_Views_on_the_Peopling_History_of_East_Asia_The_Story_According_to_HLA_Markers)
The ENA splits are very deep if the early exit of AMH out of Africa turns out to be true and also the pincer model makes the admixture history in East Asia more complex.
East asians have ancestries at least 60 or 70k separate. Just look at how different M and N mtdna are as well as C D and F y dna.
Delete