PDA

View Full Version : Reconciling the claimed West-African ancestry in certain Great Lakes clans



ThaYamamoto
07-23-2020, 05:41 PM
Reposting this from here (https://anthrogenica.com/showthread.php?17840-Great-Lakes(Africa)-peoples%92-23andme-results/page7). Thought maybe we could get a dialogue started. In another thread some time ago, Angoliga said that he'd heard the same notion of recent West-African ancestry from certain groups in Uganda, if I recall correctly. My own grandmother [of Musoga ancestry] claimed the same. Phenotypic observances aside, with the current samples and method its hard to substantiate such ancestry, seemingly. This is mirrored by South Asians with recent Central Asian ancestry [i.e. Uzbek] not being recognised on the G25. Would also like your thoughts on what seems to be the ability to hone in on African ancestry only when using unscaled coordinates. Here's my original post:

Messing around till my Ancestry results come through, 23andme have been a joke [personally] Hopefully Ancestry don't utilize 'smoothing' and the 100snp panel 23andme they use - they definitely will test for more SNPs than whatever tf 23andme were doing at v5 implementation. Anyway tangent aside, Ugandan Basoga ancestry but confused as to what 'Sudanese' is compensating for in what the Luo should theoretically already cover?

Target: yams
Distance: 2.7557% / 0.02755748
87.8 GenericBaloch+LowcasteIndian+Iran
8.2 Luo
3.0 Sudanese
1.0 Yoruba

Laka is always the closest non-Eurasian (?) admixed group to me if you remove Kikuyus etc. Also the preponderance of Cameroon populations over much closer geographic and ancestral groups (Luo+Luhya). May it be that Basoga harbor some Sahelian/Chadic ancestry or some Jola/Atlantic Coast component ?


Distance to: yams
0.02952582 GenericBaloch+LowcasteIndian+Iran
0.08127006 Laka
0.08338266 Cameroon_Bafut
0.08347043 Luo
0.08396819 Cameroon_Mbo
0.08434548 Cameroon_Aghem
0.08655513 Cameroon_Bakoko
0.08851395 Luhya_Kenya
0.09026861 Igbo
0.09043733 Sandawe
0.09216566 Yoruba
0.10099268 Hadza
0.10283138 Sudanese
0.10640659 Ethiopian_Anuak

Second edit: Has anyone noticed that you can hone in on specific ancestry (specifically african v non african) using unscaled coords? This is crazy:

Target:yams
Distance: 7.6888% / 0.07688837
70.4 Angolan
26.8 Dinka
1.6 Mbuti
1.2 Khomani_San

That is more or less exactly what you'd expect from a 100% Great Lakes Nilo-Bantu i.e. Luhya...for some reason Kikongo from Angola is always preferred but I need to message david about this. Without Angola:

Target: yams
Distance: 7.7580% / 0.07758020
35.6 Cameroon_Bangwa
22.6 Cameroon_Bafut
22.0 Dinka
15.0 Luo
2.6 Khomani_San
2.2 Mbuti

I can't use Hadza unfortunately cuz their strange psuedo-Eurasian signal messes everything up.

Here's the only study where Basoga [thankfully] been studied...notice the Guinean/Ivorian/Mende signal present clearly in the Soga and not the geographic neighboring Luhya:

38655

Kinda in keeping with this, from the new Bangime paper, we can see that the authors employed a new method (increasing K over the amount of populations involved) to reconcile Dogon populations to appear correctly instead of the usual Khoisan+Eurasian they exhibit due to an artifact effect. In this same paper, Luhya+Luos are exhibiting a lot more of the 'Atlantic' signal than I've seen in any other study where Bantu is differentiated from Atlantic. Being able to properly differentiate these signals is still in infancy, due to what I believe is the massive structurally complex profile of most African populations.

Utilizing SAPDA and a measurement of shared genetic drift specifically via rare SNPs, we see:

38656 38657

Here Dilawer/Kurd used Luhya (I erroneously presumed it was Dinka) to represent East Africa, and Yoruba+Mende for West-Africa, obviously. The Luhya are very similar to Ugandan Bantu populations East of Lake Albert, on the surface, and even the Luhya languages are phylogenetically close to Ugandan zone languages i.e. Luganda/Lusoga, part of a Nyanza continuum. Theoretically, all of the ~18% shared drift above should fall into Luhya, but we can observe that a strong Yoruba/Mende drift is being exhibited. This further helps to support such claims of recent WA ancestry in my opinion.

Finally, although my 23andme relatives list is quite short, I've recorded African-American, Afro-Cuban, Dominican and Puerto Rican matches - I'm not convinced that any peoples were directly enslaved from the Great Lakes region to the Americas, with this Eastern cohort making way to Arabia and beyond only. This is touched upon in this thesis (https://pdfs.semanticscholar.org/c0a2/939b789f2aea9c712621086f26c195612592.pdf) where the author remains unsure as to why some mtdna lineages in Uganda share exact haplotypes with African Americans/Dominicans/P-Ricans, as opposed to more 'native' Bantu haplotypes of the same clade. In Mosaic maternal ancestry in the Great Lakes region of East Africa (https://pubmed.ncbi.nlm.nih.gov/26188410/) the authors find the same phenomena, of certain clades [including Nilotic/Sudanic] ones not conforming to typical Bantu haplotypes. To me, this points to the oral history that I've been told, wherein wives from 'the West of Africa' were captured/traded/acquired some how, brought most likely through the Savannah corridor instead of the Congo basin, perhaps may be true.

Here's some final haplotype figures indicating various haplotypes of the same clade found in Uganda:

3865838659


I expect my Ancestry results and tha supporting G25/SAPDA etc to be even more indicative of this WA shift but I'll have to wait and see.

ThaYamamoto
07-23-2020, 05:59 PM
Forgot to add the admixture run from the new Bangime/Bangande paper:

38660

ThaYamamoto
07-23-2020, 08:50 PM
Been running through with individual samples this time. But the massive spanner in the works is that my paternal side also claims 'mchotara' origins, and he sure looks it, albeit through his great-great-grandmother [tribe unknown, most likely Luo but really its up in tha air].

Runs with individuals:

Target: yams
Distance: 2.8779% / 0.02877864
85.2 Baloch+LowcasteIndian+WestIran
6.2 Luo
4.6 Cameroon_Aghem
4.0 Dinka

Minus Cameroon individuals:

Target: yams
Distance: 2.8825% / 0.02882527
85.4 Baloch+LowcasteIndian+WestIran
8.6 Luo
3.8 Dinka
2.2 Esan_Nigeria

Minus 'Dinka3':

Target: yams
Distance: 2.8881% / 0.02888062
85.6 Baloch+LowcasteIndian+WestIran
9.8 Luo
2.4 Ethiopian_Anuak
2.0 Esan_Nigeria
0.2 Sudanese


Minus Anuak individuals:

Target: yams
Distance: 2.8885% / 0.02888526
85.6 Baloch+LowcasteIndian+WestIran
9.2 Luo
3.2 Sudanese
2.0 Esan_Nigeria


Returning the Cameroonians:

Target: yams
Distance: 2.8846% / 0.02884577
85.2 Baloch+LowcasteIndian+WestIran
6.6 Luo
4.4 Cameroon_Aghem
3.8 Sudanese



Distance to: yams
0.03140306 Baloch+LowcasteIndian+WestIran
0.08126075 Cameroon_Bafut:TK-KVBAF045
0.08127121 Cameroon_Mbo:CAM-MBO-083
0.08137752 Laka:T_Laka-1.DG
0.08150245 Luo:Ayodo_423C
0.08160974 Laka:T_Laka-2.DG
0.08202085 Cameroon_Bafut:TK-KVBAF012
0.08204487 Sudanese:SUDANESE3
0.08208806 Ngumba:T_Ngumba-1.DG
0.08221521 Angolan
0.08221521 AngolanAverage
0.08223430 Luo:Ayodo_502C
0.08224464 Mada:T_Mada-2.DG
0.08227527 Lemande:S_Lemande-2.DG
0.08240856 Cameroon_Aghem:TK-KVWUM025
0.08255556 Cameroon_Aghem:TK-KVWUM028
0.08256173 Kaba:T_Kaba-2.DG
0.08267140 Tikar_South:T_Tikar_South-2.DG
0.08280109 Cameroon_Aghem:TK-KVWUM023
0.08283701 Cameroon_Mbo:CAM-MBO-093
0.08293244 Bantu_S.W.:KSP201
0.08303590 Cameroon_Aghem:TK-KVWUM020
0.08321184 Cameroon_Aghem:TK-KVWUM008

The pull toward NW Cameroon populations straddling the border between Cameroon and Nigerian Calabar is interesting. Don't know what to make of Laka unless they have Eurasian ancestry I haven't pried out.

Angoliga
07-23-2020, 10:07 PM
Just to clarify, I've heard notions of recent West-African migrations into Uganda, more often than not from Nigeria oddly enough, but they're all unfounded and lack any uni-paternal data to suggest anything following initial Bantu-migrations.

Regarding the closer affinity towards Laka and other Cameroonian pops, it could be their forager-related ancestry that's causing them to gravitate towards your KEN_LSA-like forager ancestry.

If I'm not mistaken, the only notable West-African migrations into East-Africa would be across the Sahel in modern day Sudan/S.Sudan from Fulani pastoral-related groups and pops from predominantly Muslim West-Sudanic regions like the Hausa.

There's limitation using G25 to distill minor ancestry definitively due to recombination. In your case, your West-African-like affinities could descend from diverse sources (Luhya, Luo, Kikuyu) via several ancestors going back multiple generations. The alleles from that minor ancestry keep getting reshuffled every generation, so what's being rendered in G25 using modern samples, as an end product of several generations shouldn't be interpreted literally.

Out of curiosity, do you have multiple siblings who've also tested with G25? For illustrative purposes, you could run their cords using the same source samples -- you'll notice results for minor ancestry will vary due to recombination

- I hope that makes sense

ThaYamamoto
07-23-2020, 10:36 PM
Just to clarify, I've heard notions of recent West-African migrations into Uganda, more often than not from Nigeria oddly enough, but they're all unfounded and lack any uni-paternal data to suggest anything following initial Bantu-migrations.

Yeah I totally get its anecdotal and hasn't been any where near confirmed, its just weird how I've heard Nigeria too.


Regarding the closer affinity towards Laka and other Cameroonian pops, it could be their forager-related ancestry that's causing them to gravitate towards your KEN_LSA-like forager ancestry.

I see, but wouldn't Luo/Luhya and the other Bantu pops in the list have that LSA ancestry already embedded, and being far closer to my actual ancestry, trump Cameroon/Laka/Angola?


If I'm not mistaken, the only notable West-African migrations into East-Africa would be across the Sahel in modern day Sudan/S.Sudan from Fulani pastoral-related groups and pops from predominantly Muslim West-Sudanic regions like the Hausa.


For sure, that's the only confirmed migration that I know of.


There's limitation using G25 to distill minor ancestry definitively due to recombination. In your case, your West-African-like affinities could descend from diverse sources (Luhya, Luo, Kikuyu) via several ancestors going back multiple generations. The alleles from that minor ancestry keep getting reshuffled every generation, so what's being rendered in G25 using modern samples, as an end product of several generations shouldn't be interpreted literally.

Right, recombination and inheritance not my forte at all. Wouldn't the Yoruba signal inherited through that multigenerational admixture still be embedded in the Luhya sample as can be seen on SAPDA? I take your word for it though, its just this itch I'm trynna get to tha bottom of...G25 wise I understand but shared drift on SAPDA is what's nagging me. I do get that the Cameroon+Laka and Angolan samples do harbor significantly more Atlantic ancestry than Great Lakes peoples, and due to recombination like you're suggesting, that + forager = closer distance? Makes sense I gotta say.


Out of curiosity, do you have multiple siblings who've also tested with G25? For illustrative purposes, you could run their cords using the same source samples -- you'll notice results for minor ancestry will vary due to recombination

- I hope that makes sense

Nah unfortunately not yet, I'm trynna get my sister to get tested and see wassup wit that. Hopefully this COVID storm blows over by next year and I can visit Iganga with some tests aswell... It makes sense alright and I appreciate your insight man. But what would finally explain my matches in the Americas? Could thousands of years of Eastern Bantu separation still show up as IBD[?] The thesis I linked above also couldn't understand why specific mtdna haplotypes matched with AAs and Dominicans instead of those more local, and suggests possibly that the Swahili/Arabs traded with the Europeans, but I haven't been able to find anything that confirms that, and it seems very dubious.

ThaYamamoto
07-24-2020, 02:58 PM
Personal anecdotes aside heres an fst chart from the Mulindiwa paper. Notice UBB (Basoga) exhibiting more closeness with the CIV/Ivorian samples than the Luhya, and surprisingly, far closer than the Zambian sample who at least on admixture, show more westerly affinity.

38687

ThaYamamoto
07-27-2020, 02:40 AM
So turns out I have more than 50 African-American matches, and counting, on Ancestry. I can post screenshots editing names out if needed. I'm sure this amount of shared IBD further heavily supports a common West-African ancestor [within the last 400 years?] unless all of these diverse AAs (from communities across the traditional AA sphere, Louisiana/Texas/Alabama/Mississippi/Virginia/Carolinas) have Eastern Bantu ancestry which I find absolutely unlikely. Like c'mon. Do any fully Eastern Bantu/Nilotic/Sudanic people share any AA matches?

ThaYamamoto
07-28-2020, 12:35 AM
So turns out I have more than 50 African-American matches, and counting, on Ancestry. I can post screenshots editing names out if needed. I'm sure this amount of shared IBD further heavily supports a common West-African ancestor [within the last 400 years?] unless all of these diverse AAs (from communities across the traditional AA sphere, Louisiana/Texas/Alabama/Mississippi/Virginia/Carolinas) have Eastern Bantu ancestry which I find absolutely unlikely. Like c'mon. Do any fully Eastern Bantu/Nilotic/Sudanic people share any AA matches?

So no one got an opinon on tha AA matches then....right....

drobbah
07-28-2020, 01:18 AM
So no one got an opinon on tha AA matches then....right....
My mother on ancestry has around 30 matches with Arabs predominantly from Yemen & the Gulf and she has very distant Yemeni ancestry.If they are matching with you then they probably have some sort of Bantu ancestry.If I recall even some African Americans descend from SE Africans, the famous Wayan brothers have an Austronesnian haplogroup which they probably got from a Malagasy ancestor

ThaYamamoto
07-28-2020, 01:51 AM
My mother on ancestry has around 30 matches with Arabs predominantly from Yemen & the Gulf and she has very distant Yemeni ancestry.If they are matching with you then they probably have some sort of Bantu ancestry.If I recall even some African Americans descend from SE Africans, the famous Wayan brothers have an Austronesnian haplogroup which they probably got from a Malagasy ancestor

That's interesting, but as she does indeed have Yemeni ancestry I can understand, it would be what, in the last 300-400 years right? Please correct if I'm wrong I think tha consensus here is usually like 'here's mans hypothesis so now we gotta reject or confirm' but I'm just trynna conduct an inquiry..this oral history I've been passed down gotta be rooted in something, particularly cuz AAs do seem to be majority Nigerian (where I've been told we share ancestry) at this point. The small amount of Malgasy ancestry in AAs is pretty interesting too I just got an issue wit the fact that all these AA matches are from various communities/hotspots and that the Bantu ancestry would be of the western stream which would make sense on admixture but this is IBD so its like...a common ancestor from way back at Bantu dispersal would be 3-5k years ago do you know what I mean? I've still been counting and its 80+ at this point with 500 matches total so its crazy disproportionate. I haven't seen any full-blooded SE folks matches so I really don't know if I can cross reference someone from Zim for example, I do have a Zim match but no shared AA matches with him, so if anyone would like to share what they know that would be cool especially if they from Great Lakes or adjacent. Also on't wanna beat a dead horse but that thesis found shared haplotypes with AAs/Dominicans that weren't local to Uganda.

drobbah
07-28-2020, 02:11 AM
That's interesting, but as she does indeed have Yemeni ancestry I can understand, it would be what, in the last 300-400 years right? Her grandmother was mixed, which means her great grandmother was the pureblood Yemeni.Even though it's distant she still gets significant amount of matches not just from the Arabian Peninsula but even a few from the Maghreb & the Levant.


Please correct if I'm wrong I think tha consensus here is usually like 'here's mans hypothesis so now we gotta reject or confirm' but I'm just trynna conduct an inquiry..this oral history I've been passed down gotta be rooted in something, particularly cuz AAs do seem to be majority Nigerian (where I've been told we share ancestry) at this point. The small amount of Malgasy ancestry in AAs is pretty interesting too I just got an issue wit the fact that all these AA matches are from various communities/hotspots and that the Bantu ancestry would be of the western stream which would make sense on admixture but this is IBD so its like...a common ancestor from way back at Bantu dispersal would be 3-5k years ago do you know what I mean? I've still been counting and its 80+ at this point with 500 matches total so its crazy disproportionate. I haven't seen any full-blooded SE folks matches so I really don't know if I can cross reference someone from Zim for example, I do have a Zim match but no shared AA matches with him, so if anyone would like to share what they know that would be cool especially if they from Great Lakes or adjacent. Also on't wanna beat a dead horse but that thesis found shared haplotypes with AAs/Dominicans that weren't local to Uganda. Honestly, I think you should ask some of your AA matches if they have full blooded Bantu matches from the SW or the SE.I don't think that 80+ matches is a coincidence.

Deftextra
07-28-2020, 02:14 AM
So turns out I have more than 50 African-American matches, and counting, on Ancestry. I can post screenshots editing names out if needed. I'm sure this amount of shared IBD further heavily supports a common West-African ancestor [within the last 400 years?] unless all of these diverse AAs (from communities across the traditional AA sphere, Louisiana/Texas/Alabama/Mississippi/Virginia/Carolinas) have Eastern Bantu ancestry which I find absolutely unlikely. Like c'mon. Do any fully Eastern Bantu/Nilotic/Sudanic people share any AA matches?

How large are the segments for these matches? I have hundreds of matches on ancestry from all over the America for segments <13cm. But I believe this is likely because of the disproportionate amount of people testing there and most (if not all) are likely IBS segments.

ThaYamamoto
07-28-2020, 02:50 AM
Her grandmother was mixed, which means her great grandmother was the pureblood Yemeni.Even though it's distant she still gets significant amount of matches not just from the Arabian Peninsula but even a few from the Maghreb & the Levant.

Honestly, I think you should ask some of your AA matches if they have full blooded Bantu matches from the SW or the SE.I don't think that 80+ matches is a coincidence.

Right makes sense in the time frame. Cool stuff. And yeah I might just do that, I just found a 100% West African match and asked her where she from, this is an interesting development.


How large are the segments for these matches? I have hundreds of matches on ancestry from all over the America for segments <13cm. But I believe this is likely because of the disproportionate amount of people testing there and most (if not all) are likely IBS segments.

Okay now this is cool, they are all smaller segments yeah...a bunch of Mexicans and Colombians for me too lol...so I see why its iffy to base a conclusion on this just yet. Do you have straight up West African matches? I just found one waiting on what she has to say.

Deftextra
07-28-2020, 01:32 PM
Do you have straight up West African matches? I just found one waiting on what she has to say.

I have not come across individuals who are full west African yet. I have come across a few Kenyans and Tanzanians though, and from their names, its likely most of them are from the Swahili sphere. Allot of them tend to score minor amounts of south & west-Asian.

SWAHILLI_PRINCE16
07-28-2020, 02:17 PM
I have not come across individuals who are full west African yet. I have come across a few Kenyans and Tanzanians though, and from their names, its likely most of them are from the Swahili sphere. Allot of them tend to score minor amounts of south & west-Asian.

Just curious, how much West Asian & South Asian % do your Swahili matches normally score?

ThaYamamoto
07-28-2020, 03:10 PM
Just curious, how much West Asian & South Asian % do your Swahili matches normally score?

Here's some results from folks that identify as mswahili, can't find the detailed results though unfortunately.

38779

38781


Anyone able to figure out what this West-African match of mine is? Her proportions are strange...possibly Ijaw/Niger-Delta. Her name gives nothing away but her results are weird like the one Ijaw sample I've seen. Is Fonte Felipe still active on here?

Eastern Bantu Peoples: 2%
Cameroon/Congo/Southern Bantu Peoples: 52%
Nigeria: 14%
Benin & Togo: 28%
Mali: 4%

gihanga.rwanda
07-28-2020, 05:38 PM
Just curious, how much West Asian & South Asian % do your Swahili matches normally score?

Quick question. To what extent do people who identify as Swahili self segregate and differentiate themselves from each other? I am talking about Swahili people from the same town or locale. I am trying to reconcile the results of individual Swahilis, like the ones posted here, who may exhibit more cosmopolitan ancestry and the results of the following study that found only minor ancestry from SW or South Asia in 3 Swahili communities from Kenya (the Comorian group from Anjoun were the only surveyed Swahili related group with significant MENA ancestry). I know that groups in Somalia who may speak Somali as a first language but exhibit more cosmopolitan ancestry due to historic trade links are distinguished from ethnic Somalis who are homogeneous. Is this also the case along the Swahili coast? I imagine that the “Arabs” who were massacred in Zanzibar in the 60s had been there for generations, mixed with the local population, and probably spoke Swahili as a first or second language?

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5777450/

SWAHILLI_PRINCE16
07-28-2020, 07:25 PM
Quick question. To what extent do people who identify as Swahili self segregate and differentiate themselves from each other? I am talking about Swahili people from the same town or locale. I am trying to reconcile the results of individual Swahilis, like the ones posted here, who may exhibit more cosmopolitan ancestry and the results of the following study that found only minor ancestry from SW or South Asia in 3 Swahili communities from Kenya (the Comorian group from Anjoun were the only surveyed Swahili related group with significant MENA ancestry). I know that groups in Somalia who may speak Somali as a first language but exhibit more cosmopolitan ancestry due to historic trade links are distinguished from ethnic Somalis who are homogeneous. Is this also the case along the Swahili coast? I imagine that the “Arabs” who were massacred in Zanzibar in the 60s had been there for generations, mixed with the local population, and probably spoke Swahili as a first or second language?

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5777450/

There are many Swahili Sub-tribes tribes like the Mvita, Jomvu, Shirazi, Bajuni, Ngazidja of the Comoros and Changamwe to name and many others. They are all Swahili people culturally and genetically we all share a common ancestor.

The main difference ls that the tribes have different Swahili dialects
Theres the KiMvita dialect of Mombasa
The Ki Amu (Bajuni) dialect of Lamu
The Unguja Swahili dialect of Zanzibar

As for Middle Eastern admixture, some tribes might have more significant Arab+ Indian and Persian ancestry like the Bajuni & the Comorians. The admixture tends to be more recent ( Yemeni, Indian, Persian& Omani settlers marrying into the Swahili community ). There are ofcourse more pure (>95% Bantu) Swahili people with minor outside admixture.

Yes just like the Barawa of coastal Somalia who might vary in admixture compared to the other groups within the same locile, there is a similar case of admixtures varying on various coastal towns>> for example the Bajuni migrants from Lamu who live in Mombasa tend to be very mixed and look very mixed compared to the indeginous Swahili tribes of Mombasa.

Deftextra
07-28-2020, 08:18 PM
Here's some results from folks that identify as mswahili, can't find the detailed results though unfortunately.

38779

38781


Anyone able to figure out what this West-African match of mine is? Her proportions are strange...possibly Ijaw/Niger-Delta. Her name gives nothing away but her results are weird like the one Ijaw sample I've seen. Is Fonte Felipe still active on here?

Eastern Bantu Peoples: 2%
Cameroon/Congo/Southern Bantu Peoples: 52%
Nigeria: 14%
Benin & Togo: 28%
Mali: 4%

The guy from Lamu looks almost identical to one of my uncles on my fathers side lol. I would not have know the difference if someone gave me that picture.

gihanga.rwanda
07-28-2020, 08:50 PM
There are many Swahili Sub-tribes tribes like the Mvita, Jomvu, Shirazi, Bajuni, Ngazidja of the Comoros and Changamwe to name and many others. They are all Swahili people culturally and genetically we all share a common ancestor.

The main difference ls that the tribes have different Swahili dialects
Theres the KiMvita dialect of Mombasa
The Ki Amu (Bajuni) dialect of Lamu
The Unguja Swahili dialect of Zanzibar

As for Middle Eastern admixture, some tribes might have more significant Arab+ Indian and Persian ancestry like the Bajuni & the Comorians. The admixture tends to be more recent ( Yemeni, Indian, Persian& Omani settlers marrying into the Swahili community ). There are ofcourse more pure (>95% Bantu) Swahili people with minor outside admixture.

Yes just like the Barawa of coastal Somalia who might vary in admixture compared to the other groups within the same locile, there is a similar case of admixtures varying on various coastal towns>> for example the Bajuni migrants from Lamu who live in Mombasa tend to be very mixed and look very mixed compared to the indeginous Swahili tribes of Mombasa.

Thank you for this useful information! I think you’re right about the age of some of this gene flow . I wonder if the Indian/Pakistani/Persian ancestry is older than a lot of the Peninsular Arab ancestry, which might have been mediated by Oman.

In the study I linked, the Arabian ancestry in Anjouan is dated to around the time of Oman’s ascendance as a regional power. It’s interesting that the Swahili from other Comorian islands lacked this same admixture. The Comorians are a very interesting group; I wonder if they received historical Malagasy ancestry or were originally Malagasy but later Swahilized? Not surprisingly, Bantu speakers are suspected to have island hoped from the mainland through the Comoros to Madagascar.

Comorians from Anjouan result from a second admixture event between a Swahili group (Swahili from Mombasa: 74%) and a Middle Eastern group (UAE Dubai Arab: 26%) around 459 YBP (95% CI: 35–673 YBP).

SWAHILLI_PRINCE16
07-29-2020, 09:59 AM
Deleted

SWAHILLI_PRINCE16
07-29-2020, 10:12 AM
Thank you for this useful information! I think you’re right about the age of some of this gene flow . I wonder if the Indian/Pakistani/Persian ancestry is older than a lot of the Peninsular Arab ancestry, which might have been mediated by Oman.

In the study I linked, the Arabian ancestry in Anjouan is dated to around the time of Oman’s ascendance as a regional power. It’s interesting that the Swahili from other Comorian islands lacked this same admixture. The Comorians are a very interesting group; I wonder if they received historical Malagasy ancestry or were originally Malagasy but later Swahilized? Not surprisingly, Bantu speakers are suspected to have island hoped from the mainland through the Comoros to Madagascar.

Comorians from Anjouan result from a second admixture event between a Swahili group (Swahili from Mombasa: 74%) and a Middle Eastern group (UAE Dubai Arab: 26%) around 459 YBP (95% CI: 35–673 YBP).

38795

This chart might be of your interest

Source >>
https://www.youtube.com/watch?v=7YQ1gn8t3FE
The Persian & Arab admixture is older because of historical trade links, the South Asian is probably more recent. However, I could be wrong I'm basing this on the historical timeline.

ThaYamamoto
07-30-2020, 02:13 AM
Genetic drift analysis via Ancestry

3880538806

Increased from a max of 18% (23andme) to 21% (Ancestry). Combined should be interesting cuz some homozygous allele sharing between myself and Yoruba/WA are missing in the Ancestry file. Interestingly, the max range of African ancestry via Ancestrydna is 18%, in keeping with the original 23andme SAPDA analysis. Likelihood=/=reality in my case, where Ancestry give me a likelihood of 8%, but still the range feature is a lot better than 23andme. S/o to them for keeping that feature. Its just grating cuz max likelihood erases my father's tropical ancestry.

Some distances from the new file. Made a Luo/Yoruba ghost just to see:

Distance to: yammington
0.07785235 yorubaluo
0.08007734 Laka:T_Laka-1.DG
0.08034544 Cameroon_Bafut:TK-KVBAF045
0.08060099 Cameroon_Bafut:TK-KVBAF012
0.08067125 Ngumba:T_Ngumba-1.DG
0.08069157 Cameroon_Mbo:CAM-MBO-083
0.08075927 COG_Kindoki_230BP:KIN004
0.08079301 Laka:T_Laka-2.DG
0.08093417 Lemande:S_Lemande-2.DG
0.08103931 Luo:Ayodo_423C
0.08104092 Bantu_S.W.:KSP203
0.08104480 Mada:T_Mada-2.DG
0.08104622 Bantu_S.W.:KSP201
0.08116489 Kongo:S_Kongo-2
0.08120948 Tikar_South:T_Tikar_South-2.DG
0.08122192 Luo:Ayodo_502C
0.08135625 Cameroon_Aghem:TK-KVWUM020
0.08135865 Cameroon_Aghem:TK-KVWUM025
0.08137782 Cameroon_Mbo:CAM-MBO-093
0.08142248 Kaba:T_Kaba-2.DG
0.08156010 Kaba:T_Kaba-1.DG
0.08160025 Cameroon_Aghem:TK-KVWUM028
0.08170312 Bantu_S.W.:KSP198
0.08181180 Luo:Ayodo_438C
0.08197872 Cameroon_Aghem:TK-KVWUM008
0.08227138 Lemande:S_Lemande-1.DG
0.08444762 Cameroon_Bafut:TK-KVBAF081
0.08447728 Cameroon_Mbo:CAM-MBO-040
0.08454703 Cameroon_Mbo:CAM-PYKL-047
0.08460437 Cameroon_Aghem:TK-KVWUM005
0.08463173 Cameroon_Mbo:CAM-MBO-010
0.08467414 Cameroon_Aghem:TK-KVWUM026
0.08474586 Cameroon_Aghem:TK-KVWUM011
0.08476220 Cameroon_Mbo:CMF-BAK-022
0.08478372 Mende_Sierra_Leone:HG03052
0.08482458 Cameroon_Aghem:TK-KVWUM007
0.08492609 Mandenka:HGDP00912

ThaYamamoto
08-01-2020, 09:42 PM
Reminder for self to update when I have time - will edit.

ThaYamamoto
09-12-2020, 02:02 AM
As this has become a kinda personal thread I've been working on some 3D PCAs this tha only one I got saved right now:
39551

ThaYamamoto
09-14-2020, 02:05 AM
Aite so the default angle, a 2D angle and default at 1,2,4. I did get a great vertical by flipping the PCs but it crashed. Seems too intensive for my laptop. All individuals except for those tagged are the Egyptian individuals from the G25 - I didn't want to use high ANA folks for obvious reasons except user Amhas who should be 25-27% tropical African from my calculations.

https://i.imgur.com/WFP5HcN.png

https://i.imgur.com/0X1LdVY.png

https://i.imgur.com/woWDf5O.png

If anyone wants to recreate it for validity hmu and I'll send tha code.

drobbah
09-14-2020, 02:34 AM
Aite so the default angle, a 2D angle and default at 1,2,4. I did get a great vertical by flipping the PCs but it crashed. Seems too intensive for my laptop. All individuals except for those tagged are the Egyptian individuals from the G25 - I didn't want to use high ANA folks for obvious reasons except user Amhas who should be 25-27% tropical African from my calculations.

https://i.imgur.com/WFP5HcN.png

https://i.imgur.com/0X1LdVY.png

https://i.imgur.com/woWDf5O.png

If anyone wants to recreate it for validity hmu and I'll send tha code.

There's no way Amhas has 20-27% African ancestry.Looking at the way he plots, he's probably in the 13-18% range depending on the admixture model.

ThaYamamoto
09-14-2020, 03:23 AM
There's no way Amhas has 20-27% African ancestry.Looking at the way he plots, he's probably in the 13-18% range depending on the admixture model.

How you figure? 47% of Iberomaurusian=ANA=SSA-like particularly on PCAs for the time being, we know this from tha Shum Laka paper. Amhas 11% Yoruba/Dinka, 30% Iberomaurusian.. 15+11=26. If he was 13% he would't plot near the Egyptian-Intermediate profile. Lol.

drobbah
09-14-2020, 03:29 AM
How you figure? 47% of Iberomaurusian=ANA=SSA-like particularly on PCAs for the time being, we know this from tha Shum Laka paper. Amhas 11% Yoruba/Dinka, 30% Iberomaurusian.. 15+11=26. If he was 13% he would't plot near the Egyptian-Intermediate profile. Lol.
2/3 of Taforalt is Natufian-like with only 33% being ANA from what I've read.I seen his results on another thread, he's 30% IBM and 11% SSA which is roughly 20% not 25-27%.

IBM also plot closer to North Africans than the Eritreans who are around the 60% Natufian/Arabian ancestry mark.I highly doubt Taforalt and Mar_EN are 47% ANA

ThaYamamoto
09-14-2020, 04:01 AM
2/3 of Taforalt is Natufian-like with only 33% being ANA from what I've read.I seen his results on another thread, he's 30% IBM and 11% SSA which is roughly 20% not 25-27%.

IBM also plot closer to North Africans than the Eritreans who are around the 60% Natufian/Arabian ancestry mark.I highly doubt Taforalt and Mar_EN are 47% ANA

:noidea: Ask K33/Kale/Memonics what they think. Amhas and other Maghrebis will plot past any non-ANA harbouring individuals of the same Yoruba/Dinka score. Way past actually.

edit: the Shum Laka paper actually had Ghost ANA as 54% and the original Taforalt paper at 45. C'mon bro.

https://i.imgur.com/q5ainsb.jpg


46% Non-african related Taforalt. Lol.

drobbah
09-14-2020, 04:44 AM
:noidea: Ask K33/Kale/Memonics what they think. Amhas and other Maghrebis will plot past any non-ANA harbouring individuals of the same Yoruba/Dinka score. Way past actually.
From what I seen on pca plots, that's not true.For example compare Egyptian:5AJ13 which has 26% SSA ancestry with the Moroccan:MCA39 which according to my calculation would be 25% African (with yours much higher) and they plot exactly where they are supposed to plot just on a different cline.




edit: the Shum Laka paper actually had Ghost ANA as 54% and the original Taforalt paper at 45. C'mon bro.

https://i.imgur.com/q5ainsb.jpg


46% Non-african related Taforalt. Lol.

"We chose outgroups that can distinguish sub-Saharan African, Natufian, and Paleolithic European ancestries but are blind to differences between sub-Saharan African lineages (11). A two-way admixture model, comprising Natufian and sub-Saharan African populations, does not significantly deviate from our data (χ2 P ≥ 0.128), with 63.5% Natufian and 36.5% sub-Saharan African ancestry, on average (table S8)."

https://science.sciencemag.org/content/sci/360/6388/548/F4.large.jpg?width=800&height=600&carousel=1

gihanga.rwanda
09-14-2020, 03:10 PM
From what I seen on pca plots, that's not true.For example compare Egyptian:5AJ13 which has 26% SSA ancestry with the Moroccan:MCA39 which according to my calculation would be 25% African (with yours much higher) and they plot exactly where they are supposed to plot just on a different cline.

"We chose outgroups that can distinguish sub-Saharan African, Natufian, and Paleolithic European ancestries but are blind to differences between sub-Saharan African lineages (11). A two-way admixture model, comprising Natufian and sub-Saharan African populations, does not significantly deviate from our data (χ2 P ≥ 0.128), with 63.5% Natufian and 36.5% sub-Saharan African ancestry, on average (table S8)."

https://science.sciencemag.org/content/sci/360/6388/548/F4.large.jpg?width=800&height=600&carousel=1

Taforalt was modeled as approximately 65% Natufian and 35% modern SSA (Yoruba + Hadza) in the original paper by Loosdrecht et al. 2018, but this was never a great fit for Taforalt, which is 10s of thousands years older than Natufian or any of these extant populations. Lazaridis et al. 2018 proposed an alternative model using more ancient aDNA from Western Eurasia (WE) and was able to model Taforalt as approximately 55% Dzudzuana and 45% Ancient North African (ANA); the Natufians were modeled as 73% Dzudzuana and 27% Taforalt, and thus about 12% ANA, in the same paper. Lipson et al. 2020 was more focused on ancient population substructure in Africa rather than WE but build off the model proposed by Lazaridis et al. 2018. In this paper, Taforalt can be modeled as 54% ANA and 47% WE, compared to the Agaw, who can be modeled as 51% East African agro-pastoralist, 8% Mota (which contains ghost modern ancestry), and 41% WE.

These are all competing models and we won’t be able to draw any strong conclusions until we have more aDNA from North and NE Africa, but the Reich lab appears to be leaning towards a version or combination of the latter two models.

Side note. Natufians having North African ancestry is very likely IMHO and would answer a lot of lingering questions including why Natufian tends to absorb or depress some purported African ancestry in many extant populations.

ThaYamamoto
09-14-2020, 04:41 PM
From what I seen on pca plots, that's not true.For example compare Egyptian:5AJ13 which has 26% SSA ancestry with the Moroccan:MCA39 which according to my calculation would be 25% African (with yours much higher) and they plot exactly where they are supposed to plot just on a different cline.




"We chose outgroups that can distinguish sub-Saharan African, Natufian, and Paleolithic European ancestries but are blind to differences between sub-Saharan African lineages (11). A two-way admixture model, comprising Natufian and sub-Saharan African populations, does not significantly deviate from our data (χ2 P ≥ 0.128), with 63.5% Natufian and 36.5% sub-Saharan African ancestry, on average (table S8)."

https://science.sciencemag.org/content/sci/360/6388/548/F4.large.jpg?width=800&height=600&carousel=1

See what gihanga said. Your source is an earlier paper, the Shum Laka study from Reich espoused a greater proportion of ANA/SSA in Taforalt. And that's what they going with for the time being. You can already see what I linked, figures from the earlier study aren't gonna change the fact they [Reich lab] are leaning toward the 54% margin. I don't see what's controversial about this, K33's calculator had Taforalt at 53% ANA - his work is exceptional and commendable. The great thing with the Shum Laka paper was they could attempt to emulate Basal West+Central African ancestries, further increasing ANA in Taforalt which has been increasing steadily with each study.

drobbah
09-17-2020, 02:37 AM
If you are going to calculate the ANA from their IBM scores and consider it SSA, would you not have to take into consideration tthe Taforalt admixture in both Yoruba and Dinkas? They both carry Eurasian ancestry so it would still mean your exact calculations of African vs West Eurasian/Basal Eurasian in Maghrebis would be off

Mnemonics
09-17-2020, 04:40 AM
If you are going to calculate the ANA from their IBM scores and consider it SSA, would you not have to take into consideration the Taforalt admixture in both Yoruba and Dinkas? They both carry Eurasian ancestry so it would still mean your exact calculations of African vs West Eurasian/Basal Eurasian in Maghrebis would be off


I would be cautious about reading too much into the position of a specific North African. Many North Africans have high amounts of Anatolia_N admixture which can interfere with a direct comparison with Levantine rich populations. Anatolian populations had far less Basal/African affinity than Natufian/Levantine rich populations and will therefore plot further north by default. You would have to compare individuals with similar amounts of Anatolian and Levantine ancestry to judge the impact that actual ANA ancestry has on their position.


Minor note, whatever the Eurasian element in the Dinka sample from the Reichlabs dataset seems like it is not primarily from a Taforalt like population, it has some broadly Eurasian affinities that make it hard to link it to a solely West Eurasian origin.

drobbah
09-17-2020, 06:54 AM
I would be cautious about reading too much into the position of a specific North African. Many North Africans have high amounts of Anatolia_N admixture which can interfere with a direct comparison with Levantine rich populations. Anatolian populations had far less Basal/African affinity than Natufian/Levantine rich populations and will therefore plot further north by default. You would have to compare individuals with similar amounts of Anatolian and Levantine ancestry to judge the impact that actual ANA ancestry has on their position.


Minor note, whatever the Eurasian element in the Dinka sample from the Reichlabs dataset seems like it is not primarily from a Taforalt like population, it has some broadly Eurasian affinities that make it hard to link it to a solely West Eurasian origin.
If Taforalt is as both GihangeRwanda & Yamomoto reference is only 55% West Eurasian, then why do the IAM samples plot closer to Eurasians on a global pca plot than the Eritreans including the outlier sample? The Eritrean outlier sample is from what I seen 60% Natufian & Neolithic Iranian

Mnemonics
09-17-2020, 03:25 PM
If Taforalt is as both GihangeRwanda & Yamomoto reference is only 55% West Eurasian, then why do the IAM samples plot closer to Eurasians on a global pca plot than the Eritreans including the outlier sample? The Eritrean outlier sample is from what I seen 60% Natufian & Neolithic Iranian

I'm uncertain as to the exact percentage of West Eurasian ancestry in Taforalt, my qpF4ratios support the idea that they are around 55% percent Eurasian but I have had a hard time modeling them in qpadm as more than 40% Dinka.

Nevertheless, the southern position of the Eritreans can be explained as a result of the fact that both the Natufians and Neolithic Iranians are very high in Basal/African ancestry and are likely much more south shifted than the Hunter-Gatherer population that contributed to the Iberomaurusians which seems to be more rich in Anatolia_HG like ancestry.

This is evident in the same f4 ratios that show Taforalt as 55% Eurasian which indicate that the Natufians and Ganj_Dareh samples are in the 82-88 range, while Anatolia_N and Pınarbaşı are more in the 92-97 range.

gihanga.rwanda
09-17-2020, 04:21 PM
If Taforalt is as both GihangeRwanda & Yamomoto reference is only 55% West Eurasian, then why do the IAM samples plot closer to Eurasians on a global pca plot than the Eritreans including the outlier sample? The Eritrean outlier sample is from what I seen 60% Natufian & Neolithic Iranian

I think that we can consider several factors to try to answer your question.

1. Taforalt can be modeled as ~50% ANA and Dzudzuana, which the Reich lab views as a appropriate stand in for Western Eurasian-related ancestry. This model implies that Natufian has North African ancestry related to ANA or maybe one of the other “Neo-African” components described in the Shum Laka paper. If that’s the case, we’re currently underestimating the amount of African ancestry in the Middle East and adjacent regions. If Natufian can be modeled as 86% Dzudzuana and 14% ANA, Eritreans may be closer to 50% African.

2. I am also not confident that Dinka is a one to one stand in for the African ancestry in Cushitic and related groups such as Ethio-Semitic speakers. Dinka is already more West African-shifted than other extant and purported ancient Nilo-Saharan speaking groups like the Mursi etc. It also wouldn’t surprise me if Nilotic-related groups acquired more basal African ancestry after their ancestors split with the African ancestors of Cushitic-related groups. I think this is evident if you compare the yDNA/mtDNA profiles of Nilotic and Cushitic-related groups; Dinka and related groups have a lot of lineages such as yDNA B-M60 that are absent or rare in the HoA; Porto-Nilotic speakers might have picked up these lineages from indigenous hunter-gatherers when they arrived in the vicinity South Sudan escaping the drying Sahara.

3. That being said, Cushitic and related groups probably have minor basal African ancestry that is absent in Taforalt. This ancestry could’ve been present in their “Sudanese” ancestors or they might have acquired it by assimilating Mota-related groups. Any basal African ancestry, no matter how small, is going to pull you away from the Eurasian pole, which would further differentiate Taforalt from Cushitic-related groups such as Eritreans.

This is all speculative of course but based on the latest research out of the Reich lab.

ThaYamamoto
09-17-2020, 05:18 PM
If you are going to calculate the ANA from their IBM scores and consider it SSA, would you not have to take into consideration tthe Taforalt admixture in both Yoruba and Dinkas? They both carry Eurasian ancestry so it would still mean your exact calculations of African vs West Eurasian/Basal Eurasian in Maghrebis would be off

Mnemonics and gihanga etc have summed it up p much better than I can but I get what you're saying man, but I originally posted the PCAs specifically in the context of where I plot - roughly ascertaining the ancestry of individuals clustering close to me then became the imperative. Of course nothing is exact until the relevant samples are found and sequenced; however thanks to basal simulations you can roughly work out ancestral streams quite easily. In K33's thread, you can see Morocco North score the respective BasalWest+Central+East+PrePastoral+UstIshim that coincides exactly with the amount of Yoruba/Dinka scored in a standard 'deep' model plus 17-18% ANA which shifts them toward SSA (I never said ANA=SSA, but it definitely plots in this manner) thus Morocco North plots well ahead of the Egyptian average.

https://i.imgur.com/depOZr3.png


As I don't got any IBM/ANA, it was important to me to be able to figure out the max SSA-pull in these individuals/averages. What I do have is high basal/Iran Mesolithic which brings me on to my next point.

Its crucial to differentiate between G25 PCA (which is what it essentially is) and a rudimentary script like nMonte. nMonte is notorious for depressing SSA ancestral streams in high basal individuals i.e. Natufian heavy pops. I'm also very inclined to believe this same phenomenon occurs with high Iran mesolothic/neolithics i.e. Makranis and Iranian Bandaris. Awale mentions this in an older post relating to Egyptians here
(https://anthrogenica.com/showthread.php?12493-Egyptian-nMonte-modeling&p=336066&viewfull=1#post336066)

This is why PCAs can be more informative in many ways and over-reliance on nMonte can be a mistake imho. I believe I plot where a min.max 18% SSA (thanks AncestryDNA) non-IBM harboring, high Iran Mesolithic individual should when taking into account Maghrebi's SSA+ANA, Egyptians high basal resulting in SSA depression etc, and discounting nMonte maximum likelihood faults.

Mnemonics
09-17-2020, 06:20 PM
I think that we can consider several factors to try to answer your question.

1. Taforalt can be modeled as ~50% ANA and Dzudzuana, which the Reich lab views as a appropriate stand in for Western Eurasian-related ancestry. This model implies that Natufian has North African ancestry related to ANA or maybe one of the other “Neo-African” components described in the Shum Laka paper. If that’s the case, we’re currently underestimating the amount of African ancestry in the Middle East and adjacent regions. If Natufian can be modeled as 86% Dzudzuana and 14% ANA, Eritreans may be closer to 50% African.

2. I am also not confident that Dinka is a one to one stand in for the African ancestry in Cushitic and related groups such as Ethio-Semitic speakers. Dinka is already more West African-shifted than other extant and purported ancient Nilo-Saharan speaking groups like the Mursi etc. It also wouldn’t surprise me if Nilotic-related groups acquired more basal African ancestry after their ancestors split with the African ancestors of Cushitic-related groups. I think this is evident if you compare the yDNA/mtDNA profiles of Nilotic and Cushitic-related groups; Dinka and related groups have a lot of lineages such as yDNA B-M60 that are absent or rare in the HoA; Porto-Nilotic speakers might have picked up these lineages from indigenous hunter-gatherers when they arrived in the vicinity South Sudan escaping the drying Sahara.

3. That being said, Cushitic and related groups probably have minor basal African ancestry that is absent in Taforalt. This ancestry could’ve been present in their “Sudanese” ancestors or they might have acquired it by assimilating Mota-related groups. Any basal African ancestry, no matter how small, is going to pull you away from the Eurasian pole, which would further differentiate Taforalt from Cushitic-related groups such as Eritreans.

This is all speculative of course but based on the latest research out of the Reich lab.

You maybe on to something with Dinka having something more basal, there is some affinities shared by the Dinka and and Luo samples with ZAF_2000BP that seem to be less present in some West Africans.

However, I think this same basal African ancestry was also present in the African ancestors of Cushitic/Ethiosemite speakers as the same trends show up in the f4 ratios. When I try putting deeper South African groups into the F4 ratio the gap between West and East Africans in their shared ancestry with Eurasians shrinks relative to the direct comparisons, which could imply that AEA rich populations could have picked up something South African like, this also could be a result of increased affinity due to pastoralist and hunter-gatherer admixture from East Africa into South Africa, but I am uncertain about how I could adequately test the direction of the gene flow without a more pristinely AEA ancient sample.

drobbah
09-17-2020, 09:43 PM
In K33's thread, you can see Morocco North score the respective BasalWest+Central+East+PrePastoral+UstIshim that coincides exactly with the amount of Yoruba/Dinka scored in a standard 'deep' model plus 17-18% ANA which shifts them toward SSA (I never said ANA=SSA, but it definitely plots in this manner) thus Morocco North plots well ahead of the Egyptian average.



Pre-Pastoralist component that he created contains West Eurasian ancestry (also significant ANA ancestry).Isn't Ust Ishim a Eurasian sample that was prior to the differentiation of West & East Eurasian? I don't know how that correlates with the amount of Yoruba & sometimes Dinka in Maghrebis

Egyptian average plots a bit ahead of the Morocco_North sample and I ran the pca using the components in K33's G25 calculator:

https://i.imgur.com/4bcbnHH.png


I don't think we can take K33's G25 results for face value especially when results like this make no sense.Dinka being 36% Eurasian is just absurd to think about
Target: Drobbah_scaled
Distance: 6.9484% / 0.06948401
35.0 Basal_East_African
25.0 RUS_Ust_Ishim
24.8 TUR_Barcin_N
10.2 Ancestral_N_African
2.6 Caspian
2.4 WHG

Target: Dinka
Distance: 3.4691% / 0.03469119
53.8 Basal_East_African
36.0 RUS_Ust_Ishim
8.8 Basal_West_African
1.4 Ancestral_N_African










This is why PCAs can be more informative in many ways and over-reliance on nMonte can be a mistake imho. I believe I plot where a min.max 18% SSA (thanks AncestryDNA) non-IBM harboring, high Iran Mesolithic individual should when taking into account Maghrebi's SSA+ANA, Egyptians high basal resulting in SSA depression etc, and discounting nMonte maximum likelihood faults. How do we know for sure SSA suppression is happening in MENA + Horners because of their Basal heavy Natufian ancestry?

Granary
09-18-2020, 12:58 AM
The discussion is still mostly reasonable, but I wonder how far can we push the Out-of-Africa vs In-Africa admixture dichotomy deep in prehistory before it becomes worthless insofar as describing genetic relatedness between modern populations.

I think we can all agree on the basic premise that out-of-Africa is a more coherent population and an the in-Africa groupings is essentially a catch-all of deeper lineages, but let's postulate that we have X population in 40k BP Africa ends up mixing with a Out-of-Africa back migration and is 50/50 African/non-African, does it make sense to classify such a population as 50/50 in a late Paleolithic context around 10k BP considering all the drift that happened since?

At the end of the day it's arbitrary, but I wonder what people's opinions are, because when discussing things everyone has to put the line somewhere on the sand to be able to use certain terminology.

ThaYamamoto
09-18-2020, 02:11 AM
Pre-Pastoralist component that he created contains West Eurasian ancestry (also significant ANA ancestry).Isn't Ust Ishim a Eurasian sample that was prior to the differentiation of West & East Eurasian? I don't know how that correlates with the amount of Yoruba & sometimes Dinka in Maghrebis

Egyptian average plots a bit ahead of the Morocco_North sample and I ran the pca using the components in K33's G25 calculator:

https://i.imgur.com/4bcbnHH.png


I don't think we can take K33's G25 results for face value especially when results like this make no sense.Dinka being 36% Eurasian is just absurd to think about
Target: Drobbah_scaled
Distance: 6.9484% / 0.06948401
35.0 Basal_East_African
25.0 RUS_Ust_Ishim
24.8 TUR_Barcin_N
10.2 Ancestral_N_African
2.6 Caspian
2.4 WHG

Target: Dinka
Distance: 3.4691% / 0.03469119
53.8 Basal_East_African
36.0 RUS_Ust_Ishim
8.8 Basal_West_African
1.4 Ancestral_N_African








How do we know for sure SSA suppression is happening in MENA + Horners because of their Basal heavy Natufian ancestry?

The experimental K33 model is not appropriate or suitable to employ as a PCA, he's using simulations and experimental components hence tha name. So I can't comment on your PCA. I was talking specifically regarding my all Africa G25 PCA. Egyptian Average will not plot ahead of Morocco North avg on an all-Africa cline in a solid PCA, the only ANA Egyptians harbor would be nested within their Natufian ancestry if it comes to light that Natufians indeed harbor ANA.

Did you at least look at the first nMonte run I posted? Ust Ishim in this particular model is hitting 40+% in Bantu pops representing a ghost ancestral component in modern day SSAs. In the experimental calculator Ust Ishim routinely contributes toward Bantus in particular, more so than in west-west Africans, so I'm assuming here and here alone, its a proxy for something yet unknown in most modern SSAs. Its not really a problem within the confines of this model. Actually I swear K33 has discussed what Ust Ishim is playing here, I just can't be bothered going through the thread rn. Maybe check it out.The role Ust Ishim plays in general I don't think has been concluded yet, but again that isn't an interest of mine at this current time. In that same thread K33 explained what's going on with Dinka, so I'm not going to go over that again when the calculator's author has addressed it.

I'm not exactly sure as to SSA suppression in Horners cuz that isn't my field of interest, but it does frequently occur in MENA/high basal Eurasian modern groups -an artefact of nMonte stat script. Why Basal is acting this way, Mnemonics already broke it down. As did gihanga I think when they broke down the conceptualization of Basal Eurasian with each iteration/novel model per the studies.

drobbah
09-18-2020, 02:41 AM
The experimental K33 model is not appropriate or suitable to employ as a PCA, he's using simulations and experimental components hence tha name. So I can't comment on your PCA. I was talking specifically regarding my all Africa G25 PCA. Egyptian Average will not plot ahead of Morocco North avg on an all-Africa cline in a solid PCA, the only ANA Egyptians harbor would be nested within their Natufian ancestry if it comes to light that Natufians indeed harbor ANA.

Don't understand why you are so invested against Egyptians plotting closer to Eurasian especially Levantines/Arabians compared to the Maghreb and haven't seen any evidence refuting that either.You are entitled to your opinion tho


Did you at least look at the first nMonte run I posted? Ust Ishim in this particular model is hitting 40+% in Bantu pops representing a ghost ancestral component in modern day SSAs. In the experimental calculator Ust Ishim routinely contributes toward Bantus in particular, more so than in west-west Africans, so I'm assuming here and here alone, its a proxy for something yet unknown in most modern SSAs. Its not really a problem within the confines of this model. Actually I swear K33 has discussed what Ust Ishim is playing here, I just can't be bothered going through the thread rn. Maybe check it out.The role Ust Ishim plays in general I don't think has been concluded yet, but again that isn't an interest of mine at this current time. In that same thread K33 explained what's going on with Dinka, so I'm not going to go over that again when the calculator's author has addressed it.

I'm not exactly sure as to SSA suppression in Horners cuz that isn't my field of interest, but it does frequently occur in MENA/high basal Eurasian modern groups -an artefact of nMonte stat script. Why Basal is acting this way, Mnemonics already broke it down. As did gihanga I think when they broke down the conceptualization of Basal Eurasian with each iteration/novel model per the studies.
All you said in the above just goes to show how confused everyone is about African genetics.So untill we get ancient African genomes all you or anyone else can do is shoot in the dark and hypothesize.Anyways continue on with the thread!



The discussion is still mostly reasonable, but I wonder how far can we push the Out-of-Africa vs In-Africa admixture dichotomy deep in prehistory before it becomes worthless insofar as describing genetic relatedness between modern populations.

I think we can all agree on the basic premise that out-of-Africa is a more coherent population and an the in-Africa groupings is essentially a catch-all of deeper lineages, but let's postulate that we have X population in 40k BP Africa ends up mixing with a Out-of-Africa back migration and is 50/50 African/non-African, does it make sense to classify such a population as 50/50 in a late Paleolithic context around 10k BP considering all the drift that happened since?

At the end of the day it's arbitrary, but I wonder what people's opinions are, because when discussing things everyone has to put the line somewhere on the sand to be able to use certain terminology.
It becomes an issue when one attempts to use the dichotomy and estimate the "Africanness" and non-African percentages of a population without any ancient genomes from the Sahara for AEA,Western Africa,the Horn, Arabia (possible Basal Eurasian sample) and most importantly for me anyways, the Nile Valley & Red Sea Hills.

ThaYamamoto
09-18-2020, 03:11 AM
Don't understand why you are so invested against Egyptians plotting closer to Eurasian especially Levantines/Arabians compared to the Maghreb and haven't seen any evidence refuting that either.You are entitled to your opinion tho


All you said in the above just goes to show how confused everyone is about African genetics.So untill we get ancient African genomes all you or anyone else can do is shoot in the dark and hypothesize.Anyways continue on with the thread!

Lol so you don't got an actual riposte to none of the points you been going around in circles like NASCAR from myself or any of the more knowledgeable members...right...

drobbah
09-18-2020, 03:26 AM
Lol so you don't got an actual riposte to none of the points you been going around in circles like NASCAR from myself or any of the more knowledgeable members...right...
Just to clarify, you haven't provided any concrete evidence that the Northern Moroccan samples plot closer to Eurasians (Levantines,Arabians) than the Egyptian samples which is the main reason why I entered this thread in the first place.Anyways this thread is about Bantus, and I shouldn't waste your time so I'll leave you to it

ThaYamamoto
09-18-2020, 03:41 AM
Just to clarify, you haven't provided any concrete evidence that the Northern Moroccan samples plot closer to Eurasians (Levantines,Arabians) than the Egyptian samples which is the main reason why I entered this thread in the first place.Anyways this thread is about Bantus, and I shouldn't waste your time so I'll leave you to it

Man do you even read what's posted? The whole point is the Egyptian average is closer to Eurasians than Maghrebis precisely because IBM shifts Maghrebis further toward SSA. I've said this countless times, the Egyptian average/non outliers are closer to Eurasians. Why that was relevant is in relation to me on the appropriate Africa PCA.

edit: think I just got it. When I said plots 'ahead' you think I mean toward Eurasian, but what I actually mean is toward SSA. So Maghreb_North ahead of Egyptian to me means more SSA. Is this the misunderstanding going on here?