Page 1 of 2 12 LastLast
Results 1 to 10 of 18

Thread: ADMIXTURE is precient again: the Ethio-Somali component and the Pastoral Neolithic

  1. #1
    Registered Users
    Posts
    1,472
    Sex

    ADMIXTURE is precient again: the Ethio-Somali component and the Pastoral Neolithic

    People give ADMIXTURE a lot of shit, but honestly its pretty good at finding clines of admixture the peaks of which subsequently get confirmed in aDNA studies (e.g
    Lithuanians as peaking in EHG, Sardinians are EEF-rich, Georgians are CHG-rich, each cline and peak have a real basis in gradients of admixture formed by ancient populations) as long as you don't reify components (e.g. EHG = Lithuanians! or Lithuanian-centered component in Eurogenes k12 is 'real' EHG and better than Dodecad's Lithuanian-centered component!) or take drift-centered components as real (e.g. Palestinians/Kalash/Ari cultivators are ancient populations come to life!).

    In Hodgson et. al's 2014 study we see an "Ethio-Somali" component that generated a lot of attention and excitement. You can see its distribution as the dark green component in fig 2. It peaks in Somalis and other Cushitic speakers of the Horn and is well represented in Ethiosemitic speakers, pastoralists of Kenya and Tanzania such as Maasai and Luhya, and appears in South African Khoikhoi pastoralists as well.

    When it appears in Somalis, it eats up all of the "Arabian" component in them (which probably forms its bulk), all of the "Ari blacksmith" component, and a lot of the "Nilo-saharan" component. The Somalis become ~85% "Ethio-Somali" with the rest Nilo-saharan. Notably, the Cushites of Ethiopia are likewise extremely "Ethio-Somali" with little or no "Arabian" ancestry left over, but the Ethio-Semites are much less so with quite a bit of the "Arabian" ancestry in them not eaten up by this component.

    The patterns here seem really clear to me: from the three-part composition of this component ("Arabian"::Natufian, "Ari Blacksmith"::Mota, "Nilo-saharan"::Sudan-relsted ancestry for which we have no ancient samples yet) to the distribution (peaks in Cushites but casts a long shadow into East Africa down to Khoisan populations), this component seems like a dead ringer for the Pastoral Neolithic genomes, and seems to trace ancient Cushitic-speaking populations just like PN did. Yet another goal for this good ol' algoritm!

    The more interesting implications come with the distribution of this component in other Afroasiatics. This component is significantly present among all Semitic and Berber populations as well, including in Eurasia and NW Africa. It does not displace the "Bantu/West African" signal, showing that the phenomena are probably separate. The fact that its found in NW Africa is especially interesting given some members' interest in signals of E African admixture in Berbers. It seems like with Dinka the signal is slight to nonexistent, but here we have something a little different, but still interesting, especially in that this component occurs in levels in NW Africa inconsistent with introduction from the Middle East/Arab populations as its only vector.

    ADMIXTURE is not a formal admixture test, which means you cannot be sure that components represent the same thing everywhere they occur. So we cannot say "Ethio-Somali component in Somalis (presumably Natufian + Mota + Dinka) = same thing in Berbers!". The most we can say is that the patterns of allele frequency sharing among Cushites are significantly correlated with patterns among Berbers and Semitic speakers, and probably Afroasiatics as a whole.

    While AA is a very old language family and we may doubt that it may even have left a discernible population-genetic signal, I think this is a pretty good indication that in fact it did, with the signal detectable even using only modern populations, and with the PN genomes we are tantalizingly close. Even if the ratios of the three ancestries have to be shifted a bit, something like them is probably not too far from the real thing.
    Last edited by Ryukendo; 11-27-2020 at 06:51 AM.
    Quoted from this Forum:

    "Which superman haplogroup is the toughest - R1a or R1b? And which SNP mutation spoke Indo-European first? There's only one way for us to find out ... fight!"

    " A Basal Eurasian and an Aurignacian walk into a bar... "

    " No, you are in the wrong... I really hope that you are not jumping on my thread with intent to harass me, just like other "receiving comitee", that unites in classic bullying unity, which makes me sad about such people, deprived of love etc.... "

  2. The Following 13 Users Say Thank You to Ryukendo For This Useful Post:

     alexfritz (11-29-2020),  Alfa (06-21-2021),  beyoku (11-27-2020),  blackflash16 (11-28-2020),  Brandon S. Pilcher (11-27-2020),  drobbah (11-27-2020),  Gadzooks (11-27-2020),  gihanga.rwanda (12-06-2020),  Lenny Nero (11-28-2020),  Megalophias (11-27-2020),  pgbk87 (11-28-2020),  siberoberingian (11-29-2020),  ThaYamamoto (11-27-2020)

  3. #2
    Registered Users
    Posts
    243

    I think this is worth looking at, however I can't help but feel that outside of Sub Saharan Africa, the Ethio-Somali component is not going to represent Mota and Dinka like ancestry themselves but something related from the East African branch, like ANA, that is likely to be the population associated with Proto Afroasiatic.

    The Ethio Somali component also turns up in Greece and Southern Anatolia where as far as I know it's quite difficult to find traces of post-ANA African ancestry. I could be mistaken about this though.

  4. The Following 2 Users Say Thank You to Psynome For This Useful Post:

     davit (11-29-2020),  pgbk87 (11-29-2020)

  5. #3
    Banned
    Posts
    672
    Sex
    Ethnicity
    1/2 Italian, 1/2 Armenian
    Nationality
    USA
    Y-DNA (P)
    R1b-U152
    mtDNA (M)
    H5a

    Quote Originally Posted by Psynome View Post
    I think this is worth looking at, however I can't help but feel that outside of Sub Saharan Africa, the Ethio-Somali component is not going to represent Mota and Dinka like ancestry themselves but something related from the East African branch, like ANA, that is likely to be the population associated with Proto Afroasiatic.

    The Ethio Somali component also turns up in Greece and Southern Anatolia where as far as I know it's quite difficult to find traces of post-ANA African ancestry. I could be mistaken about this though.
    How much of this component turns up in those places?

  6. #4
    Registered Users
    Posts
    243

    Quote Originally Posted by davit View Post
    How much of this component turns up in those places?
    If you click the link to the article in the OP and go to the ADMIXTURE figures with maps, you'll see the light green Ethio-Somali component extend into Greece and Anatolia. Looking at the bar graphs you can see trace amounts in a variety of European populations at the highest K value.

    It's quite small but clearly visible, and intermediate regions often don't display it. This suggests to me that it may be picking up on something very old, but I'm not sure if it's telling us there's something actually Ethio-Somali like there. Rather it may be picking up on an ancestry source shared between Ethio-Somalis and Europeans.

    With ADMIXTURE it's difficult to tell exactly what's happening, which is one reason why it's less favored for use in analyses lately.

  7. #5
    Registered Users
    Posts
    121
    Sex
    Y-DNA (P)
    R1a-M417
    mtDNA (M)
    U2b2

    United States of America India
    You mean dark green? Light is north africa? Could be related to basal eurasian / deep ancestry. It’s higher in north africa, arabia, and southern europe. Lesser amounts in europe and west asia. And then absent in south asia

  8. #6
    Registered Users
    Posts
    1,472
    Sex

    Yep dark green.

    Quote Originally Posted by Psynome View Post
    If you click the link to the article in the OP and go to the ADMIXTURE figures with maps, you'll see the light green Ethio-Somali component extend into Greece and Anatolia. Looking at the bar graphs you can see trace amounts in a variety of European populations at the highest K value.

    It's quite small but clearly visible, and intermediate regions often don't display it. This suggests to me that it may be picking up on something very old, but I'm not sure if it's telling us there's something actually Ethio-Somali like there. Rather it may be picking up on an ancestry source shared between Ethio-Somalis and Europeans.

    With ADMIXTURE it's difficult to tell exactly what's happening, which is one reason why it's less favored for use in analyses lately.
    If you look at the maps, there's actually no datapoints in Anatolia, Greece or southern Spain, the green color there is purely due to interpolation between places where samples do exist.
    Quoted from this Forum:

    "Which superman haplogroup is the toughest - R1a or R1b? And which SNP mutation spoke Indo-European first? There's only one way for us to find out ... fight!"

    " A Basal Eurasian and an Aurignacian walk into a bar... "

    " No, you are in the wrong... I really hope that you are not jumping on my thread with intent to harass me, just like other "receiving comitee", that unites in classic bullying unity, which makes me sad about such people, deprived of love etc.... "

  9. The Following User Says Thank You to Ryukendo For This Useful Post:

     Psynome (11-29-2020)

  10. #7
    Registered Users
    Posts
    243

    Quote Originally Posted by Ryukendo View Post
    Yep dark green.



    If you look at the maps, there's actually no datapoints in Anatolia, Greece or southern Spain, the green color there is purely due to interpolation between places where samples do exist.
    Thanks for clearing that up. How about the component apparently showing up in French Basques, Tuscans and the like? Doesn't that seem a bit off? I'm just looking at the ADMIXTURE graphs in the figure.

  11. The Following 2 Users Say Thank You to Psynome For This Useful Post:

     MethCat (06-26-2021),  Ryukendo (11-29-2020)

  12. #8
    Banned
    Posts
    955
    Sex
    Ethnicity
    Mixed Zanzibar
    Y-DNA (P)
    L1307
    mtDNA (M)
    L3b1a

    ZanzibarSultanate United Kingdom Uganda
    Yes ADMIXTURE identifies some very interesting things particularly when it comes to African populations. For example I've seen it repeatedly determine the difference between Dinka/Dinka-related peoples and the east-African side of Horn of Africans, something that has been obvious to me for quite a while but nice to see vindicated sometimes.

    F1.large (1).jpg

  13. The Following 5 Users Say Thank You to ThaYamamoto For This Useful Post:

     drobbah (11-29-2020),  gihanga.rwanda (12-06-2020),  Lenny Nero (11-29-2020),  Mansamusa (11-30-2020),  siberoberingian (11-29-2020)

  14. #9
    Banned
    Posts
    121
    Sex
    Location
    Winter is quite cold you know

    Hawaii
    No it's meaningless and outdated study without ancient data.

  15. #10
    Gold Class Member
    Posts
    701
    Sex
    Location
    Gulf of Fars
    Ethnicity
    Somali
    Y-DNA (P)
    E-V32
    mtDNA (M)
    N1a

    Somalia Ethiopia Eritrea Djibouti
    Quote Originally Posted by Ryukendo View Post
    Even if the ratios of the three ancestries have to be shifted a bit, something like them is probably not too far from the real thing.
    Crazy stuff. Lank and I theorized what you did almost a decade ago now. We both suspected that Ethio-Somali might be representing a real population which we suspected was early Cushites who were possibly more MENA than current Horners and it seems that when you use Kenyan Pastoral Neolithic samples for Horners something close to the Ethio-Somali picture appears:




    Mota related ancestry is abysmally depleted in this model because these Pastoral Neolithic samples have a lot of Mota-like ancestry but you get the gist. I suspect that when a pure Neolithic Cushite is found, meaning one that is just a straight mixture between Dinka-like people in Sudan and pre-historic Egyptians, this model will shift toward significant Mota admixture appearing across the board (5-15%) for all populations and perhaps even the extra South-Sudanese stuff disappearing as it might be showing up in Horners above to compensate for the lower amounts in the Pastoral Neolithic samples due to Mota-related admixture. Or Horners just did in fact gain real extra Nilo-Saharan admixture somehow which sorta fits with how studies showing an Ethio-Somali cluster like Hobson et al. always showed some extra Nilo-Saharan ancestry outside of the component:

     





    Either way I think the gist of Horners' history and genetic ancestry is:

    * Neolithic Sudanese people expand into the Horn around 3000 BCE and bring Cushitic, pottery, grain collecting with eventually locally developed farming, cattle, goats, sheep, donkeys, shepherd/hunting dogs, bows and arrows, rock art and carvings, stelaes, burial mounds, cairns, dolmens and about 60-80% or more of the ancestry in nearly all Cushites and Ethiosemites.

    * Gradual mixing over-time occurs in varying degrees with native Horn Hunter-Gatherers who would have resembled Mota with populations tracing anything from 5 to 20% or so of their ancestry to them.

    * Southern Arabians begin to enter the Horn around 1000 BCE and over-time bring things like Proto-Ethiosemitic, metallurgy, writing, monumental architecture, new agricultural techniques, camel domestication and admixture into the local livestock like asiatic cattle that mix with the local longhorns while contributing to anything from 5 to 30% of Horners' ancestry.

    No great changes outside of internal migrations and mixing occur after this with the Horn remaining pretty much the result of these migrations and intermixture by periods like
    when Aksum is around and the Middle-Ages. I could be wrong and some odd curveballs like the South-Sudanese being legit and metallurgy and stuff predating Arabians in some areas or not entirely coming from them could be the case (Horn archaeology really needs some work) but that's just the sense I get from the data I've seen for now.
    Last edited by Awale; 06-21-2021 at 07:03 AM.
    فار عارابإ آ واجإب اه

  16. The Following 7 Users Say Thank You to Awale For This Useful Post:

     Alfa (06-21-2021),  CopperAxe (06-21-2021),  drobbah (06-21-2021),  Meres (07-02-2021),  pgbk87 (06-23-2021),  Ryukendo (06-21-2021),  traject (06-21-2021)

Page 1 of 2 12 LastLast

Similar Threads

  1. Replies: 5
    Last Post: 08-18-2019, 05:49 AM
  2. Replies: 23
    Last Post: 11-12-2017, 03:10 PM
  3. Replies: 9
    Last Post: 09-19-2017, 04:01 PM
  4. Replies: 51
    Last Post: 01-30-2016, 06:33 AM
  5. WHG Component in Neolithic Anatolia
    By J Man in forum Autosomal (auDNA)
    Replies: 6
    Last Post: 07-03-2015, 03:43 PM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •