Page 9 of 9 FirstFirst ... 789
Results 81 to 85 of 85

Thread: Upcoming K53 Ancient Eurasia Calculator

  1. #81
    Registered Users
    Posts
    456
    Sex
    Ethnicity
    Central/E. Sudanic
    Nationality
    Canadian
    Y-DNA
    A-M13 (A-Y95250*)
    mtDNA
    L3x2b

    Canada Uganda Southern Sudan Chad Ethiopia Omo Region Sudan
    Quote Originally Posted by Kurd View Post
    African-E - Masai
    African-E2- Hadza
    African-S - Various pygmies
    African-W- Yorubans & Gambians
    Maasai pops are heavily admixed with Cushitic and Bantu affinities:

     

    (Hollfelder, 2017)

    The Dinka/Nuer/Gumuz.. would've been a more preferable choice for a "Nilotic" aka "Ancient_East_African" component; this affinity is even non-negligible among SE-Mediterraneans, Levantines and other SW-Asians so it might be worth tweaking for these Eurasian testers.
    Last edited by NiloSaharan; 06-22-2018 at 07:27 PM.
    PuntDNAL K8
    Population: Hidden Content 46.16, Ubangian_Congo 9.56, W_Benue_Congo 24.21, Eastern_HG 2.09, E_Benue_Congo 12.94, Omotic 5.05
    Single Population Sharing: Hidden Content 14.17, Hidden Content 17.26, South_Sudanese 25.21, South_Sudan_Anuak 25.65, Cameroon_Mada 28.57, Ethiopian_Gumuz 33.91, Kenyan_Maasai 36.07, Kenyan_Bantu 38.05, Chad_Kaba 40.04, DRC_Hema 42.08, Kenyan_Luhya 46.64

    The truth is not for all men, but only for those who seek it - Hidden Content

  2. The Following 5 Users Say Thank You to NiloSaharan For This Useful Post:

     Kurd (06-22-2018),  lukaszM (06-22-2018),  okarinaofsteiner (06-26-2018),  palamede (06-22-2018),  randwulf (06-22-2018)

  3. #82
    Banned
    Posts
    5,173
    Sex

    Quote Originally Posted by NiloSaharan View Post
    Maasai pops are heavily admixed with Cushitic and Bantu affinities:

     

    (Hollfelder, 2017)

    The Dinka/Nuer/Gumuz.. would've been a more preferable choice for a "Nilotic" aka "Ancient_East_African" component; this affinity is even non-negligible among SE-Mediterraneans, Levantines and other SW-Asians so it might be worth tweaking for these Eurasian testers.
    yes, but I’m limited to what I have in my data set, and more importantly marker overlap.

    The biggest handicap I face with accuracy in ancient calculators is insufficient marker overlap. This is by far my biggest obstacle and sometimes makes it impossible to achieve accuracy for all users. I wish HGDP , other public databases and companies like 23, ancestry, FTDNA genotyped a ton more markers. It would make my life a 1000 times easier and I wouldn’t have to abandon projects halfway

    To give you an example of 1 of the many problems I face with this calculator. The Baikal-HG samples onlly have 20 K overlapping markers at 150K markers. This means the tester’s genome is compared at only 15% of the positions. Thus Baikal doesn’t have a fighting chance with the other components. This can cause a few testers depending on the compant to get 98% Baikal where others from asimilar ethnic background get 1%.

    To circumvent those problems with Baikal, I would have to prune everything back to 20K SNPs. The problem with that is discrimination between closely related pops such as in the K45 is shot. The other alternative if Baikal is included would be to drop down to a K8 or so more ancient calculator.

    I can go on and on with the challenges. If only more SNPs were genotyped .......

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

     NiloSaharan (06-22-2018),  PoxVoldius (06-23-2018)

  5. #83
    Registered Users
    Posts
    456
    Sex
    Ethnicity
    Central/E. Sudanic
    Nationality
    Canadian
    Y-DNA
    A-M13 (A-Y95250*)
    mtDNA
    L3x2b

    Canada Uganda Southern Sudan Chad Ethiopia Omo Region Sudan
    Quote Originally Posted by Kurd View Post
    yes, but I’m limited to what I have in my data set, and more importantly marker overlap.

    The biggest handicap I face with accuracy in ancient calculators is insufficient marker overlap. This is by far my biggest obstacle and sometimes makes it impossible to achieve accuracy for all users. I wish HGDP , other public databases and companies like 23, ancestry, FTDNA genotyped a ton more markers. It would make my life a 1000 times easier and I wouldn’t have to abandon projects halfway

    To give you an example of 1 of the many problems I face with this calculator. The Baikal-HG samples onlly have 20 K overlapping markers at 150K markers. This means the tester’s genome is compared at only 15% of the positions. Thus Baikal doesn’t have a fighting chance with the other components. This can cause a few testers depending on the compant to get 98% Baikal where others from asimilar ethnic background get 1%.

    To circumvent those problems with Baikal, I would have to prune everything back to 20K SNPs. The problem with that is discrimination between closely related pops such as in the K45 is shot. The other alternative if Baikal is included would be to drop down to a K8 or so more ancient calculator.

    I can go on and on with the challenges. If only more SNPs were genotyped .......
    Thank you for the rundown Kurd, I appreciate you sharing these technical complexities -- I doubt the average laymen like myself would ever come across an explanation of these technical nuamces.

    Overall great job! , we can only do with what we have
    PuntDNAL K8
    Population: Hidden Content 46.16, Ubangian_Congo 9.56, W_Benue_Congo 24.21, Eastern_HG 2.09, E_Benue_Congo 12.94, Omotic 5.05
    Single Population Sharing: Hidden Content 14.17, Hidden Content 17.26, South_Sudanese 25.21, South_Sudan_Anuak 25.65, Cameroon_Mada 28.57, Ethiopian_Gumuz 33.91, Kenyan_Maasai 36.07, Kenyan_Bantu 38.05, Chad_Kaba 40.04, DRC_Hema 42.08, Kenyan_Luhya 46.64

    The truth is not for all men, but only for those who seek it - Hidden Content

  6. #84
    Registered Users
    Posts
    43
    Sex

    Hi Kurd,please I want to test with your Calculator K53,I don't know your Email for to send my raw data.Thanks

  7. #85
    Registered Users
    Posts
    12
    Sex

    Hello there I'd also like to know

Page 9 of 9 FirstFirst ... 789

Similar Threads

  1. Replies: 570
    Last Post: 08-22-2018, 07:15 PM
  2. *New* Ancient-Eurasia K10 DIY calculator
    By Passa in forum Autosomal (auDNA)
    Replies: 268
    Last Post: 08-21-2018, 11:59 PM
  3. Upcoming Ancient Eurasia K6 Calculator
    By Kurd in forum Autosomal (auDNA)
    Replies: 143
    Last Post: 05-25-2018, 09:27 AM
  4. Ancient Eurasia K20 admixture calculator
    By Kurd in forum Ancient (aDNA)
    Replies: 22
    Last Post: 10-21-2017, 12:46 PM
  5. Upcoming Eurasia ASI / CHG calculator
    By Kurd in forum Autosomal (auDNA)
    Replies: 94
    Last Post: 12-29-2015, 02:37 PM

Tags for this Thread

Posting Permissions

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