Page 8 of 8 FirstFirst ... 678
Results 71 to 79 of 79

Thread: Gene Flow Between Southern and South-eastern Asia

  1. #71
    Bronze Class Member
    Posts
    2,385
    Sex
    Location
    Krasnoyarsk
    Y-DNA (P)
    R1a-Y6
    mtDNA (M)
    U7b
    mtDNA (P)
    G3a3

    Siberian Tatars Tajikistan
    Quote Originally Posted by DMXX View Post
    Right you are - However, I haven't found a clear indication in the paper that they used both samples in their basic qpAdm model. Perhaps I overlooked it? Did you see anything regarding how they treated IVCp in their model?
    Well, no, I haven't been able to find a direct quote from the paper itself but others who've read it have basically said the same thing I have. And the thing is that if they're using strictly SIS3+Onge+Sintashta to model south Asians, it's going to result in a poor fit for a lot of South Asians particularly those originating in the northwest because that's too much AASI and not enough Iran_N. Kalash alone have similar levels of Iran_N compared to SIS3 and way less AASI.

    IVCp apparently was described as being a straight mix of Iranian farmer and AASI ancestry:

    "In fact, it was certainly the case that the peoples of the Indus Valley were genetically
    297 heterogeneous as we observe one of the Indus_Periphery individuals having ~42% AASI
    298 ancestry and the other two individuals having ~14-18% AASI ancestry (but always mixes of the
    299 same two proximal sources of AASI and Iranian agriculturalist-related ancestry)."

    There were 4 total SIS samples, 2 of which belonged to SIS1. So perhaps they've used 3 of the 4? It's quite confusing
    Last edited by Censored; 06-08-2019 at 03:36 AM.

  2. #72
    Gold Class Member
    Posts
    6,729

    Quote Originally Posted by DMXX View Post
    Right you are - However, I haven't found a clear indication in the paper that they used both samples in their basic qpAdm model. Perhaps I overlooked it? Did you see anything regarding how they treated IVCp in their model?
    "Indus_Periphery sample (a pool of three outlier individuals from the BMAC site of Gonur and from Shahr-i-Sokhta)"


    "Gonur2_BA This sample has significant amounts of South Asian HG related ancestry and is part of a genetic grouping we call Indus_Periphery"

    "The following two samples have significant amounts of South Asian HG related admixture, and we call these samples part of the Indus_Periphery genetic grouping.
    ... Shahr-i-Soktha_MLBA2 ... Shahr-i-Soktha_MLBA3"

    For each Indian Cline group X, we estimated the vector of coefficients under qpAdm:
    c(X) = (cS(X), cI(X), cA(X)) (4.1)
    where cj(X) are admixture coefficients for S=Steppe-related (Steppe_MLBA), I=Iranian
    agriculturalist-related (Indus_Periphery), and A=South Asian hunter-gatherer (AASI)-related (Onge).
    https://www.biorxiv.org/content/bior...1/292581-1.pdf
    Last edited by parasar; 06-09-2019 at 05:24 PM.

  3. The Following User Says Thank You to parasar For This Useful Post:

     bmoney (06-17-2019)

  4. #73
    Administrator
    Posts
    3,759
    Sex
    Y-DNA (P)
    R2a*-M124 (L295-)
    mtDNA (M)
    D4j5*

    England
    You guys are right, and that's been my interpretation since reading the paper when the pre-print first came out.

    The question I'm posing is with respect to the qpAdm three pop model they feature - What is being used for their IVCp ref pop? Are they using SiSBA2? SiSBA3? Both (somehow)?

    I vaguely remember reading somewhere (a comment in this forum perhaps) suggesting they used SiSBA3. That made sense prima facie, given SiSBA3 is the most AASI-shifted sample we currently have. Didn't double-check the statement concerning that and assumed it was true, but now I can't find any obvious statement in the paper regarding this, now that Censored's given me/us a valid reason to doubt that assumption is true.

  5. #74
    Registered Users
    Posts
    2,889
    Location
    Gonur Tepe

    Afghanistan Jammu and Kashmir United States of America Canada
    Quote Originally Posted by DMXX View Post
    You guys are right, and that's been my interpretation since reading the paper when the pre-print first came out.

    The question I'm posing is with respect to the qpAdm three pop model they feature - What is being used for their IVCp ref pop? Are they using SiSBA2? SiSBA3? Both (somehow)?

    I vaguely remember reading somewhere (a comment in this forum perhaps) suggesting they used SiSBA3. That made sense prima facie, given SiSBA3 is the most AASI-shifted sample we currently have. Didn't double-check the statement concerning that and assumed it was true, but now I can't find any obvious statement in the paper regarding this, now that Censored's given me/us a valid reason to doubt that assumption is true.
    Its a trio of SIS2, SIS3 and Gonur2, I told poi to simulate the averages on G25, it essentially is 70% Iran_N related/30% AASI, their criteria for Iran_N is a more Hotu/Sarazm version so it includes WSHG/AG3 like ancestry, there is 13 of them now, most are still from Shahr I Shokta and they are very likely coming from Balochistan/Sind and I don't think the main cluster is ancestral for most Indian populations they are way too Iran_N shifted, Gangetic and Insular Indians need a more tribal shifted base that is very evident from the figure below. In the new figures they are more broad , Steppe MLBA is now just Steppe (include Dali or WSHG sources I am assuming) and Iran_N is now Iran related source, its a smart way of not admitting error but fixing it at the same time. Rather the 2 outliers are most ancestral and likely coming from further east.

    I wish they included the Daamgard samples so the "ANI" side would be filled up .


    Last edited by pegasus; 06-10-2019 at 02:19 AM.

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

     bmoney (06-17-2019),  MonkeyDLuffy (06-10-2019),  Observer (06-26-2019),  okarinaofsteiner (07-29-2019),  Reza (06-10-2019),  Sapporo (06-10-2019)

  7. #75
    Registered Users
    Posts
    56
    Sex
    Omitted
    Ethnicity
    Mostly Chinese
    Nationality
    USA

    United States of America China
    Quote Originally Posted by poi View Post
    I modeled a few Southeast Asians using a few Subcontinental-heavy models:

    Here, I picked a few Steppe-rich inner South Asians like various Brahmin pops + No/Less-Steppe SouthAsians like Chamar and Velamas + Onge + Southern Chinese + Papuan:

     


    The model is pretty good for Thai and Vietnam, but very bad for Burmese and Cambodian. In this model, we're could be missing more western-type of East Asian necessary.

    So, let's add the ancient Nepal Chokhopani sample from ~750BCE:

     


    Note how adding the ancient Chokhopani brings the Burmese down to a good fit, but does not change Cambodian. Furthermore, Chokhopani helps improve Vietnam.

    Basically, Vietnam is entirely East Asian with Southern Chinese mixed with a bit of Western Chinese (proxied by Chokhopani).
    Cambodian probably need more SEAsian farmers rather than straight up Onge.

    Here, added DMXX's simulated AA... things are even better for Burmese, Thai, and Vietnam, plus Cambodian fit distance also goes down.



    Note how Nepali Brahmins are preferred over other South Asians, for Burnese and Thai, when we add the simulated AA? No Chamar, no Velamas, and not even Gangetic Brahmins. Strange right? Does that model makes sense? Just 4 references needed to model 3 out of 4 SE Asian groups. Re-running the model again by removing all those references scoring zero.



    Thoughts?

    Alright, replaced Nepal Brahmins with Rors. Things get even better for Burmese and Thais



    Basically, Ror-like pops brought along IndoAryan to Burmese and Thais. I'm curious to see if that holds up for more southern pops like Malays and Indonesians.

    I wonder if the Nepali Brahmin sample is slightly East Eurasian ("Mongoloid") admixed, since that would explain why it's a better fit for the S_Asian ancestry of those groups.

    Khmers having less actual South Asian ancestry but more "para-AASI"/Hoabinhian (Razib Khan's "Ancestral SE Asian") makes sense.

  8. The Following 2 Users Say Thank You to okarinaofsteiner For This Useful Post:

     bmoney (08-05-2019),  Kulin (07-29-2019)

  9. #76
    Registered Users
    Posts
    485
    Sex
    Ethnicity
    1/2 Ilocano 1/2 Tagalog
    Nationality
    American
    Y-DNA (P)
    C-M130
    mtDNA (M)
    B4a1a

    Philippines
    How would you interpret this 6.7% South Asian percentage I get on the LivingDNA.com test as a filipino?:


  10. The Following 3 Users Say Thank You to shazou For This Useful Post:

     okarinaofsteiner (09-13-2019),  tipirneni (09-13-2019),  Tsakhur (09-11-2019)

  11. #77
    Registered Users
    Posts
    56
    Sex
    Omitted
    Ethnicity
    Mostly Chinese
    Nationality
    USA

    United States of America China
    Quote Originally Posted by shazou View Post
    How would you interpret this 6.7% South Asian percentage I get on the LivingDNA.com test as a filipino?:

    I have your MyHeritage results in my private GEDmatch dataset. You scored 0.51% "South_Central_Asian" and 3.21% "South_Indian" in MDLP K23b, so you have some genuine South Asian ancestry. You also scored another ~2% West Eurasian, which could just as easily be Spanish as it could be Indian.

  12. The Following 2 Users Say Thank You to okarinaofsteiner For This Useful Post:

     Nibelung (09-13-2019),  shazou (09-16-2019)

  13. #78
    Registered Users
    Posts
    421
    Sex
    Location
    NYC
    Y-DNA (P)
    N-Z1927
    mtDNA (M)
    H1h
    Y-DNA (M)
    R-U152

    United States of America Denmark Norway Sweden Germany Italy
    I should like to point out that EST IA 0LS10_1 can be modeled with a full quarter Sintashta, so some trace Uralic ancestry could perhaps in the right combinations masquerade as South Asian in Europe

    fit 2.5334 pen=0

    Estonia BA 52.5
    Sintashta MLBA 25
    Corded Ware Baltic 16.67
    Nganassan 4.17
    Shamanka N 1.67
    My 23andMe kit into Eurogenes K36 then oracle (thanks to lukaszM):

    nMonte restricted: Ireland 48.05, SW-England 20.95, Finnish East 20.05, Russian Tver 3.70, Latvian 3.15, Mari 1.90, Lithuanian 1.30, French Basque 0.90, Orcadian 0.00

    nMonte full: Ireland 45.70, Finnish East 20.05, SW-England 20.05, Russian Tver 4.10, Orcadian 3.05, Latvian 2.45, Mari 1.95, Lithuanian 1.45, French Basque 1.00, W-England 0.15, Belarusian East 0.05

  14. #79
    Registered Users
    Posts
    485
    Sex
    Ethnicity
    1/2 Ilocano 1/2 Tagalog
    Nationality
    American
    Y-DNA (P)
    C-M130
    mtDNA (M)
    B4a1a

    Philippines
    Quote Originally Posted by okarinaofsteiner View Post
    I have your MyHeritage results in my private GEDmatch dataset. You scored 0.51% "South_Central_Asian" and 3.21% "South_Indian" in MDLP K23b, so you have some genuine South Asian ancestry. You also scored another ~2% West Eurasian, which could just as easily be Spanish as it could be Indian.
    my mother scored like 2.7% south asian on the recent 23andme BETA v5.2 update (she is of full tagalog ethnicity(50% from nueva ecija & 50% from bulacan region):


Page 8 of 8 FirstFirst ... 678

Similar Threads

  1. Replies: 0
    Last Post: 07-06-2017, 04:45 PM
  2. Replies: 13
    Last Post: 02-20-2016, 04:17 PM
  3. Replies: 43
    Last Post: 01-18-2015, 02:11 PM
  4. Arabian conquest and female gene flow in North Africa and Iraq
    By Alanson in forum General Sociology/Ethnology
    Replies: 9
    Last Post: 12-19-2014, 06:43 PM
  5. Third gene flow?
    By thetick in forum Human Evolution
    Replies: 1
    Last Post: 02-18-2014, 12:46 AM

Posting Permissions

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