African phylogeny (beginning...)

[dporter]

Check out the PCA G25 thread in the African subforum

[Kale]

So what happened to Chad single-handedly re-writing our understanding of human phylogenetic history?

Same thing that always happens?

In all seriousness though, I think understand.
There are many times when I've tried to do similar things that get 99% of the way there and think you're really on to something and then one little piece that doesn't fit forces you to trash it

[Michalis Moriopoulos]

It's pretty clear to me that Basal Humans, ancient Central Africans, and ancient Southern Africans existed. It's also clear that there was a Eurasian bottleneck. For everything in between Central/Southern African and the Eurasian bottleneck, I'm just content to wait for the ancient DNA rather than making any assumptions. It just seems too speculative to bother risking getting attached to any theory.

[maroco]

How much? I would say its just a glitc or you have some kind of "old ancestry" for which they don't have the right reference from Africa or Eurasia, but a similar ancestry survived among Papuans. So its not back migration from Papuans, but a similarity based on very ancient ancestry - if the algorithm made no mistake.

One percent it’s consistent in my autosomal through to my g25. Like you said similarities based on very ancient ancestry

[piye]

It's pretty clear to me that Basal Humans, ancient Central Africans, and ancient Southern Africans existed. It's also clear that there was a Eurasian bottleneck. For everything in between Central/Southern African and the Eurasian bottleneck, I'm just content to wait for the ancient DNA rather than making any assumptions. It just seems too speculative to bother risking getting attached to any theory.

Thank you very much

A point well made

[Kale]

Hey Chad, got a question for you with my very basic understanding of these graphs. I suppose the numbers that each line is labelled with represents drift. Also assuming drift is a function of time, roughly. It appears the drift between the ghost 'East Eurasia' and Tianyuan is 438. Whereas ghost 'South Africa' and South_Africa_2000BP is 152. How is it possible that the large time difference between the latter two nodes (70k - 100k BP?) is represented by roughly a third of the drift between early East Eurasian and Tianyuan?

I know this was asked a long time ago, but it doesn't look like anyone answered. In the readme file for qpgraph there is a note (paraphrasing) 'branch lengths ending at single samples are uninformative'.
The regular 1240k typed samples are 100% similar to themselves because they are pseudohaploid, there is no variation because there is only 1 set of data so you always get ginormous branch lengths adding up to 400-500.
South_Africa_2000BP is (2-4?) high quality samples. In any population there is variation so there are spots where they are different from each other, so the branch length drops.

The only single sample exceptions are .DG samples like Ust_Ishim and Loschbour. They are diploid so the two sets of chromosomes are compared to each other and differ.
Ust_Ishim especially is very interesting because wherever he is reasonably put on a graph the terminal edge is very very small.
He must not have lived very long at all after some important phylogenic branching.

[Kale]

Here's my latest attempt at African phylogeny. A few pesky 0 edges (which in various runs seem to be resolved when Taforalt takes more 'Aterian' ancestry, but it doesn't always do so for whatever reason)

 

digraph G {
size = "7.5,10" ;
labelloc = "t" ;
label = "122920g :: Din Los Men Cam -0.000330 0.000655 0.000985 0.000377 2.614
" ;

Denisova_DG [ label = "Denisova_DG" ] ;
Neanderthal_DG [ label = "Neanderthal_DG" ] ;
Mota_SG [ label = "Mota_SG" ] ;
South_Africa_2000BP_SG [ label = "South_Africa_2000BP_SG" ] ;
Taforalt [ label = "Taforalt" ] ;
Onge_DG [ label = "Onge_DG" ] ;
Mbuti_DG [ label = "Mbuti_DG" ] ;
Dinka_DG [ label = "Dinka_DG" ] ;
Mende_DG [ label = "Mende_DG" ] ;
Cameroon_ShumLaka_8000BP_DG [ label = "Cameroon_ShumLaka_8000BP_DG" ] ;
Loschbour_DG [ label = "Loschbour_DG" ] ;
Root -> Erectus [ label = "0" ] ;
Erectus -> Heidelberg [ label = "19" ] ;
Heidelberg -> A00 [ label = "58" ] ;
Heidelberg -> Neandersovan [ label = "30" ] ;
pDEN -> Denisova_DG [ label = "63" ] ;
A00 -> Aterian [ label = "3" ] ;
A00 -> AM32 [ label = "3" ] ;
Neandersovan -> pNEA [ label = "35" ] ;
AM32 -> B2b [ label = "6" ] ;
AM32 -> AM13 [ label = "18" ] ;
pNEA -> Neanderthal_DG [ label = "7" ] ;
B2b -> ASA [ label = "2" ] ;
B2b -> ACA [ label = "14" ] ;
OOA -> OOA1 [ label = "28" ] ;
IBM1 -> IBM2 [ label = "28" ] ;
AWA -> AWA1 [ label = "0" ] ;
ASA -> South_Africa_2000BP_SG [ label = "42" ] ;
ASA -> AEA [ label = "39" ] ;
OOA1 -> Onge_DG [ label = "37" ] ;
OOA1 -> W [ label = "12" ] ;
pMOT -> Mota_SG [ label = "151" ] ;
IBM2 -> Taforalt [ label = "45" ] ;
AEA -> E [ label = "0" ] ;
W -> Loschbour_DG [ label = "35" ] ;
pDIN2 -> Dinka_DG [ label = "2" ] ;
pMEN -> Mende_DG [ label = "4" ] ;
pCAM2 -> Cameroon_ShumLaka_8000BP_DG [ label = "9" ] ;
pMBU -> Mbuti_DG [ label = "11" ] ;
E -> EV38 [ label = "0" ] ;
EV38 -> EM2 [ label = "13" ] ;
Neandersovan -> pDEN [ style=dotted, label = "45%" ] ;
Erectus -> pDEN [ style=dotted, label = "55%" ] ;
AEA -> OOA [ style=dotted, label = "97%" ] ;
pNEA -> OOA [ style=dotted, label = "3%" ] ;
IBM -> IBM1 [ style=dotted, label = "98%" ] ;
Aterian -> IBM1 [ style=dotted, label = "2%" ] ;
EM2 -> AWA [ style=dotted, label = "62%" ] ;
Aterian -> AWA [ style=dotted, label = "38%" ] ;
EV38 -> pMOT [ style=dotted, label = "72%" ] ;
AM13 -> pMOT [ style=dotted, label = "28%" ] ;
EM2 -> pDIN1 [ style=dotted, label = "74%" ] ;
AM13 -> pDIN1 [ style=dotted, label = "26%" ] ;
pDIN1 -> pDIN2 [ style=dotted, label = "96%" ] ;
IBM2 -> pDIN2 [ style=dotted, label = "4%" ] ;
AWA1 -> pMEN [ style=dotted, label = "95%" ] ;
IBM2 -> pMEN [ style=dotted, label = "5%" ] ;
ACA -> pCAM1 [ style=dotted, label = "30%" ] ;
AWA1 -> pCAM1 [ style=dotted, label = "70%" ] ;
pCAM1 -> pCAM2 [ style=dotted, label = "97%" ] ;
IBM2 -> pCAM2 [ style=dotted, label = "3%" ] ;
ACA -> pMBU [ style=dotted, label = "46%" ] ;
AWA1 -> pMBU [ style=dotted, label = "54%" ] ;
E -> IBM [ style=dotted, label = "49%" ] ;
W -> IBM [ style=dotted, label = "51%" ] ;
}

[beyoku]

Bump LOL

[Chad Rohlfsen]

I'm alive. I'm sorry. The last couple years have been difficult in finding extra time raising these kids and working enough to keep ahead. I do plan on finishing everything I've started.

As for those edges, they tend to resolve themselves with new aDNA we have from the UP.

[Woozler]

Will the model account for back-to-Africa migrations ("Ancient Admixture into Africa from the ancestors of non-Africans" - "We find evidence for substantial migration from the ancestors of present-day Eurasians into African groups between 40 and 70 thousand years ago, predating the divergence of Eastern and Western Eurasian lineages")?

https://www.biorxiv.org/content/10.1...555v1.full.pdf

It's quite likely that the Levant served as the source of several back-to-Africa migrations that massively reshaped the genetic history of the entire African continent. First was the early Upper Paleolithic migration described in the article above, taking place possibly in the early stages of the Emiran archaeological culture. Modern Niger-Kordofanian and Nilo-Saharan speakers can trace a significant chunk of their ancestry to these early Levantine migrants.

The "Emirans" later started to receive substantial gene flow (and cultural influences) from the north (Europe/Anatolia), turning into "Levantine-Aurignacians". During the later stages of this culture, around 25k ybp, some of these Levantine Aurignacians migrated across the shore of North Africa to the Maghreb (by that point seemingly unpopulated for almost 20 000 years), becoming "Iberomaurusians". The Levantines, meanwhile, continued to receive gene inflows from the north, eventually becoming Natufians, who again migrated to Africa during the Holocene, forming the bulk of the ancestry of the speakers of Afro-Asiatic languages.

"Basal Eurasian", on the other hand, is a meaningless concept resulting from the facile oversimplification of UP Eurasian population genetics into an untenable "Basal/East/West Eurasian" trichotomy. There were multiple separate admixture events involving archaic humans, haplogroup C and D carriers initially had distinctly separate histories across much of the world (Kostenki/Sunghir mammoth hunters and Australian Aborigines were both initially pure C's, it seems) than the now-dominant haplogroup F and its branches, etc., etc.

Any criticism of any of the above?