McColl et al. 2018 models East Asians as roughly 75% Australasian and 25% Tianyuan, while Xia et al. 2019 models East Asians as roughly 75% Tianyuan and 25% Australasian. What could account for the discrepancy? Xia seems to have included Siberian and Steppe populations such as Devil's Gate.
https://science.sciencemag.org/content/361/6397/88#F1
https://www.biorxiv.org/content/10.1...0903v1.full#T5
Interesting question. Complex admixture graphs seem really opaque to me, there so many possible relationships that could affect the outcome. McColl et al's models include archaic humans while the Xia et al's do not (6-7% Neanderthal in main Eurasian branch is considerably more than usual estimates); McColl et al have Kostenki as outgroup to East Eurasians while Xia et al have Ust' Ishim; and they have different modern samples too. For instance, supposing East Asians (including Tianyuan) but not Australasians had an affinity to Ust' Ishim, then that might constrain high Tianyuan input in the one graph; conversely, if Tianyuan had an affinity to Kostenki 14 that other East Eurasians didn't, that could force low Tianyuan in the other graph. At least, I think that's how it would work?
McColl stated Tianyuan and East Asians formed a clad in D-stats, which seems a bit contradictory to their qpGraph tree, to my limited understanding..
Hando (12-16-2019)
A clade against who? I don't think they form a clade against Onge. Could be against Papuans, but the Denisovan could make up for that.
Last edited by Megalophias; 12-15-2019 at 08:30 PM.
Hando (12-16-2019)
You can perhaps find more information in this article: The GenomeAsia 100K Project enables genetic discoveries across Asia
On here, https://www.nature.com/articles/s415...93-z/figures/5, you can find some split times for several South Asian, American, East Asian and Oceanian populations.
There's more information in the supplementary material (supplementary note 3, Figure S3.3):
We found that the most ancient lineages in Southeast Asia and Oceania are the Melanesians and Negritos, who show substructure from ~40 kya. Within the Negrito groups, there is evidence of separation around 20 – 30 kya. The large estimated population sizes for Philippine and Malay Negrito groups from 10 – 20 kya suggests they had prosperous settlements in the region. In contrast the Andamanese show reduced population size estimates since their split from other Negrito groups, possibly due to their geographic isolation (Figure S4.2C). The separation among Indian groups gradually occurred from between 10 – 24 kya, which is consistent with previous studies (Basu et al. 2016). The split of Northeast Asians with Austronesians is estimated at 6 – 23 kya, earlier than the estimated split time between Northeast Asians and Mongolians at 12 – 16 kya. Therefore, Koreans, Japanese and Mongolians likely share recent common ancestry.
Hando (12-16-2019)
Even with Tianyuan, the Hoabinhians, Devil's Gate, Kolyma, and Jomon in tow, a solid East Eurasian phylogeny is still frustratingly evasive.
Do Australo-Melanesians, AASI, Andamanese, Semang, and Philippine Negritos constitute an "Australasian clade," even if a weak one? I'm not sure. One confounding factor is that Philippine Negritos have Austronesian ancestry and Semang (Malaysian Negritos) have Austro-Asiatic ancestry. But even in admixture runs, much of the ancestry of all Negritos (even accounting for Neolithic East Asian admixture) seems to come out as East Asian-related rather than Australo-Melanesian-related. One look at the Hoabinhian and Andamanese genomes confirms this. I wonder if it might have something to do with these groups being ancestral to East Asians in some way (and have retained some drift-sharing with them), or perhaps they have some basal East Asian admixture that Papuans lack.
Ελευθερία ή θάνατος.
Even among the onge and Jarawa ?Originally Posted by Michalis Moriopoulos
Posting Permissions
Reply With Quote