Page 58 of 58 FirstFirst ... 848565758
Results 571 to 572 of 572

Thread: Ancient DNA from North Africa (this time with formal stats)

  1. #571
    Registered Users
    Posts
    1,891
    Sex
    Omitted

    With much appreciated help from Robert Maier I now have the automated qpgraph up and running. Testing it out now with the following populations...
    Denisova.DG, Neanderthal.DG, South_Africa_2000BP.SG, Mbuti.DG, Mende.DG, Dinka.DG, Cameroon_ShumLaka_8000BP.DG, Mota.SG, Taforalt, Onge.DG, Loschbour.DG.
    I didn't give it any starter graph just to let it do it's thing, I just set it to have 12 admixture edges (what my last manual graph had 122920h I think I posted it on Chad's Africa thread) to see if it can find something better.
    122920h had a score of 31. This automated one starts wonky, up in the thousands, but I'm excitedly watching it improve, currently at 294 after 17 generations.

    First run got stuck in a local minimum of 134.8 after 46 generations. Not sure what the graph looked like I don't know how to access the best final graph when the operation is aborted.
    Giving it another run, just doing 20 generations this time to see where it gets graph wise. I'll feed it the product of that the next run.
    EDIT: After 20 generations the score was 501. Was a very janky looking graph, I'm just going to feed it 122920h for 20 generations.

    EDIT: Ok here's the first product! It was able to reduce 122920h's score of 31 down to 19.6691. It seemed to do so with 2 crucial changes.
    - Pushing the deep ancestry in West Africa to a 'GhostArchaic' rather than 'BasalHuman' sort of spot.
    - Creating a trifurcation of 1) Mota&Dinka deep ancestry, 2) Taforalt deep ancestry, 3) South-Africa/Central-Africa/OOA
    Some of the nodes have inaccurate/misleading labeling now because they got moved around.
     

    digraph G {
    size = "7.5,10";
    Root -> Erectus [ label = "0" ];
    Erectus -> Aterian [ label = "1" ];
    Erectus -> pDENa [ label = "1" ];
    Heidelberg -> A00 [ label = "3" ];
    Heidelberg -> Neandersovan [ label = "81" ];
    A00 -> AM32 [ label = "0" ];
    A00 -> IBM1a [ label = "0" ];
    Aterian -> Heidelberg [ label = "0" ];
    Aterian -> AWAa [ label = "0" ];
    Neandersovan -> pNEA [ label = "29" ];
    Neandersovan -> pDENb [ label = "34" ];
    AM32 -> B2b [ label = "5" ];
    AM32 -> AM13 [ label = "9" ];
    pNEA -> NeanderthalDG [ label = "20" ];
    pNEA -> OOAa [ label = "0" ];
    pDEN -> DenisovaDG [ label = "43" ];
    B2b -> ASA [ label = "3" ];
    B2b -> ACA [ label = "2" ];
    ASA -> South_Africa_2000BPSG [ label = "17" ];
    ASA -> AEA [ label = "51" ];
    AEA -> E [ label = "9" ];
    AEA -> OOAb [ label = "10" ];
    OOA -> OOA1 [ label = "10" ];
    OOA1 -> OngeDG [ label = "37" ];
    OOA1 -> W [ label = "12" ];
    W -> LoschbourDG [ label = "34" ];
    W -> IBMa [ label = "1" ];
    E -> E1 [ label = "1" ];
    E -> pMOTa [ label = "45" ];
    E1 -> EM2 [ label = "18" ];
    E1 -> IBMb [ label = "0" ];
    EM2 -> AWAb [ label = "0" ];
    EM2 -> pDIN1a [ label = "0" ];
    AM13 -> pMOTb [ label = "23" ];
    AM13 -> pDIN1b [ label = "0" ];
    pMOT -> MotaSG [ label = "132" ];
    pDIN2 -> DinkaDG [ label = "1" ];
    IBM1 -> IBM2_multi2_1 [ label = "3" ];
    IBM2_multi2_1 -> IBM2_multi2_2 [ label = "30" ];
    IBM2_multi2_1 -> pCAM2a [ label = "0" ];
    AWA -> AWA1_multi1_1 [ label = "0" ];
    AWA1_multi1_1 -> AWA1_multi1_2 [ label = "4" ];
    AWA1_multi1_1 -> pMENa [ label = "2" ];
    pMEN -> MendeDG [ label = "2" ];
    ACA -> pMBUa [ label = "2" ];
    ACA -> pCAM1a [ label = "0" ];
    pCAM2 -> Cameroon_ShumLaka_8000BPDG [ label = "3" ];
    pMBU -> MbutiDG [ label = "11" ];
    pDIN1 -> pDIN2a [ label = "1" ];
    IBM -> IBM1b [ label = "2" ];
    pCAM1 -> pCAM2b [ label = "3" ];
    IBM2_multi2_2 -> IBM2_multi2_3 [ label = "20" ];
    IBM2_multi2_2 -> pDIN2b [ label = "0" ];
    IBM2_multi2_3 -> Taforalt [ label = "1" ];
    IBM2_multi2_3 -> pMENb [ label = "0" ];
    AWA1_multi1_2 -> pMBUb [ label = "3" ];
    AWA1_multi1_2 -> pCAM1b [ label = "2" ];
    pDENa -> pDEN [ style=dotted, label = "11%" ];
    pDENb -> pDEN [ style=dotted, label = "89%" ];
    OOAa -> OOA [ style=dotted, label = "5%" ];
    OOAb -> OOA [ style=dotted, label = "95%" ];
    pMOTa -> pMOT [ style=dotted, label = "58%" ];
    pMOTb -> pMOT [ style=dotted, label = "42%" ];
    pDIN2a -> pDIN2 [ style=dotted, label = "97%" ];
    pDIN2b -> pDIN2 [ style=dotted, label = "3%" ];
    IBM1a -> IBM1 [ style=dotted, label = "11%" ];
    IBM1b -> IBM1 [ style=dotted, label = "89%" ];
    AWAa -> AWA [ style=dotted, label = "48%" ];
    AWAb -> AWA [ style=dotted, label = "52%" ];
    pMENa -> pMEN [ style=dotted, label = "97%" ];
    pMENb -> pMEN [ style=dotted, label = "3%" ];
    pCAM2a -> pCAM2 [ style=dotted, label = "5%" ];
    pCAM2b -> pCAM2 [ style=dotted, label = "95%" ];
    pMBUa -> pMBU [ style=dotted, label = "46%" ];
    pMBUb -> pMBU [ style=dotted, label = "54%" ];
    pDIN1a -> pDIN1 [ style=dotted, label = "60%" ];
    pDIN1b -> pDIN1 [ style=dotted, label = "40%" ];
    IBMa -> IBM [ style=dotted, label = "54%" ];
    IBMb -> IBM [ style=dotted, label = "46%" ];
    pCAM1a -> pCAM1 [ style=dotted, label = "31%" ];
    pCAM1b -> pCAM1 [ style=dotted, label = "69%" ];
    }

    EDIT: Ran that through another 20 generations, no improvement. Going to see how many admixtures can be cut without trashing it tomorrow.
    Last edited by Kale; 01-05-2021 at 06:19 PM.
    Collection of 14,000 d-stats: Hidden Content Part 2: Hidden Content Part 3: Hidden Content PM me for d-stats, qpadm, qpgraph, or f3-outgroup nmonte models.

  2. The Following 6 Users Say Thank You to Kale For This Useful Post:

     blackflash16 (01-06-2021),  durmang (01-05-2021),  Megalophias (01-05-2021),  Mnemonics (01-05-2021),  ThaYamamoto (01-06-2021),  theplayer (01-05-2021)

  3. #572
    Registered Users
    Posts
    145
    Sex

    Quote Originally Posted by gihanga.rwanda View Post
    Scroll back a few pages in this thread for Chadís models.

    https://anthrogenica.com/showthread....n-Avgs)/page87



    According to the most recent run, Taforalt can be roughly modeled as 35% Sahel (which peaks in West Africans), ~10% Horn of African (peaks in NE Africans), and 55% Western Eurasian, primarily Natufian like. This corroborates the results from Lazaridis et al. 2018.

    Chad - Feel free to jump in.
    Interesting but East African ancestry in Iran_N and Iron_Gates_HG is very unlikely. Never seen a paper use that model nor are there any haplogroups that indicate such an affinity.

Page 58 of 58 FirstFirst ... 848565758

Similar Threads

  1. Understanding Formal Statistics, f4 and D-Stats
    By TuaMan in forum Inquiries Corner
    Replies: 31
    Last Post: 08-21-2020, 01:19 PM
  2. Shotgun .vs. 1240k in formal stats
    By Kale in forum General
    Replies: 3
    Last Post: 03-16-2020, 12:09 AM
  3. Replies: 3
    Last Post: 02-08-2019, 02:30 PM
  4. Ancient Iberomaurusian mtDNA from North Africa
    By Callingstar in forum Ancient (aDNA)
    Replies: 15
    Last Post: 01-08-2017, 04:49 AM
  5. Kurd Genetics using Formal Stats
    By Kurd in forum Western
    Replies: 55
    Last Post: 05-20-2016, 06:53 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
  •