Page 2 of 2 FirstFirst 12
Results 11 to 19 of 19

Thread: Egyptian nMonte modeling

  1. #11
    Gold Member Class
    Posts
    544
    Sex
    Location
    Gulf of Fars
    Ethnicity
    Somali
    Y-DNA
    E-V32
    mtDNA
    N1a

    Somalia Ethiopia Eritrea Djibouti
    Quote Originally Posted by pegasus View Post
    Awale can u post Iran recent and Iron Age Iran results
    Sure. In another thread but sure.

  2. The Following 3 Users Say Thank You to Awale For This Useful Post:

     Agamemnon (11-10-2017), jesus (11-09-2017), Power77 (11-09-2017)

  3. #12
    Registered Users
    Posts
    643
    Sex
    Ethnicity
    Parsi Brahmin
    Y-DNA
    L-SK1414
    mtDNA
    U8b1a1

    African Union Ainu AchaemenidEmpire1 Kurdistan Star of David Dravida Nadu
    Quote Originally Posted by pegasus View Post
    Awale can u post Iran recent and Iron Age Iran results
    I remember playing around with the sample(with the sheikh Awale). They were mainly Iran ChL + 15-20% steppe.

  4. The Following 3 Users Say Thank You to jesus For This Useful Post:

     Agamemnon (11-10-2017), Awale (11-09-2017), Power77 (11-09-2017)

  5. #13
    Registered Users
    Posts
    534
    Location
    Gonur Tepe

    Quote Originally Posted by Awale View Post
    Sure. In another thread but sure.
    I just wanted to see how CHG and Levantine components increase over time

  6. The Following User Says Thank You to pegasus For This Useful Post:

     Awale (11-10-2017)

  7. #14
    Gold Class Member
    Posts
    791
    Sex
    Location
    United States
    Ethnicity
    LebaGermish
    Y-DNA
    P312>DF19>DF88>S4268
    mtDNA
    J2a1a1e

    United States of America Lebanon Germany England Belgium Scotland
    Quote Originally Posted by Awale View Post
    I'd love to but those ancient samples aren't in the Global-10 PCA (I assume there aren't enough markers, otherwise David would have them up). But I've seen people like Need4Speed meticulously pore over them via ADMIXTURE calcs over at Gedmatch and they looked very very similar to Copts which makes sense, to be honest.
    What about 3Drif-26? He was supposed to look Copt-like.
    R1b (aka M343) > M269 > L23 > L51 > L11 > P312 > DF19 > DF88 > FGC11833 > S4281 > S4268 > Z17112 (S17075-, L644-)

    Archaeological cousin: 6DRIF-23 of Driffield Terrace Roman Cemetery, York (Z17112+, S17075+, L644-)

    Known ancestors: Francis Cooke (I-M223/I2a2a) b. 1583; John Wing (U106) b. 1584; Richard Warren (M269Hidden Content ) b.c. 1578; Elizabeth Walker (Warren) (H1j mtDNA) b. 1583; John Mead b.c. 1634 (I2a1/P37.2)

  8. The Following 3 Users Say Thank You to Dewsloth For This Useful Post:

     Lank (11-10-2017), nee4speed111 (11-10-2017), Targum (11-10-2017)

  9. #15
    Registered Users
    Posts
    265
    Sex

    Egypt
    Hey Awale, I've got the raw data of another copt from 23andme if you would like it, along with something like 27 or so coptic Gedmatch kits. I'm pretty representative of copts based on the results I've seen, just slightly less Dinka shifted than the average coptic sample.

  10. The Following 3 Users Say Thank You to nee4speed111 For This Useful Post:

     Agamemnon (11-10-2017), Awale (11-10-2017), Targum (11-10-2017)

  11. #16
    Gold Member Class
    Posts
    544
    Sex
    Location
    Gulf of Fars
    Ethnicity
    Somali
    Y-DNA
    E-V32
    mtDNA
    N1a

    Somalia Ethiopia Eritrea Djibouti
    Quote Originally Posted by nee4speed111 View Post
    Hey Awale, I've got the raw data of another copt from 23andme if you would like it, along with something like 27 or so coptic Gedmatch kits. I'm pretty representative of copts based on the results I've seen, just slightly less Dinka shifted than the average coptic sample.
    Do email that Copt's raw data! Thanks.

    Quote Originally Posted by Dewsloth View Post
    What about 3Drif-26? He was supposed to look Copt-like.
    I was under the impression that he wasn't in the Global-10 PCA but I think this might be him:

    England_Roman_outlier:3DT26,0.0066,0.0308,-0.0053,0.0435,0.0061,0.012,-0.0297,0.0027,-0.0048,-0.0008

    I'll run some models soon and see if it is and try and see if he fits better with Copts or Levantines.
    Last edited by Awale; 11-10-2017 at 02:11 PM.

  12. The Following 3 Users Say Thank You to Awale For This Useful Post:

     Agamemnon (11-10-2017), Dewsloth (11-10-2017), Targum (11-10-2017)

  13. #17
    Registered Users
    Posts
    265
    Sex

    Egypt
    Quote Originally Posted by Awale View Post
    Recently sent David nee4speed111's raw data so that he could throw him into the global-10 PCA as "Egyptian-Copt" (he seems rather representative/typical of Copts, from what I've seen but but I'd no doubt prefer more samples) and here we go:

    Egyptian_Copt,0.0034,0.027,-0.0043,0.0406,0.0012,0.0108,-0.0252,-0.0043,-0.0092,-0.004

    I ran some models quickly after just for the heck of it:

    Code:
    [1] "distance%=0.9432 / distance=0.009432"
    
             Egyptian_Copt
                          
    Natufian          38.8
    Iran_ChL          27.4
    Tepecik_Ciftlik_N 26.3
    Dinka              7.5
    
    
    39% Natufian, 27% Zagros-Chl, 26% Anatolia_N and 8% Dinka.
    
    --
    
    [1] "distance%=1.1044 / distance=0.011044"
    
             Egyptian_Copt
                 
    Levant_N 46.5
    Iran_ChL 30.3
    Natufian 13.8
    Dinka     9.4
    
    47% Levant_N, 30% Zagros-Chl, 14% Natufian and 9% Dinka.
    
    --
    
    [1] "distance%=0.8286 / distance=0.008286"
    
             Egyptian_Copt
                          
    Natufian          39.6
    Tepecik_Ciftlik_N 36.9
    Iran_N            17.4
    Dinka              6.1
    
    40% Natufian, 37% Anatolia_N, 17% Zagros_N and 6% Dinka.
    
    --
    
    [1] "distance%=0.7256 / distance=0.007256"
    
             Egyptian_Copt
                                  
    Tepecik_Ciftlik_N:Tep004 35.10
    Natufian:I1072           15.90
    Natufian:I0861           13.75
    Levant_N:I1704           11.60
    Iran_N:WC1                9.60
    Iran_N:AH1                7.65
    Dinka                     6.40
    
    35% Anatolia_N, 30% Natufian, 12% Levant_N, 17% Zagros_N and 6% Dinka.
    
    --
    
    [1] "distance%=2.9821 / distance=0.029821"
    
             Egyptian_Copt
                     
    Saharawi     69.0
    Satsurblia   31.1
    
    69% Saharawi and 31% CHG.
    I don't know what but something's missing for Copts, it seems. Something holding models like the first one from being really good fits (i.e. below or at around 0.5), even using individual samples instead of population averages (for the ancients) like with the 4th model didn't help all that much. It's odd because using the very same samples (+Yorubas) as the 1st model actually produced a decent fit for the usual non-Copt Egyptian samples out there:

    [1] "distance%=0.3783 / distance=0.003783"

    Egyptian

    Natufian 33.80
    Iran_ChL 28.85
    Tepecik_Ciftlik_N 24.40
    Dinka 9.65
    Yoruba 3.30

    I ran some other models for non-Copt Egyptians as well and I'd say the first and last one below are probably the most sensible ones:

    Code:
    [1] "distance%=0.5151 / distance=0.005151"
    
             Egyptian
                           
    Egyptian_Copt      77.0
    Palestinian        16.4
    Mende_Sierra_Leone  4.1
    Bantu_SA_Tswana     2.5
    Arab_Israel_1       0.0
    Arab_Israel_2       0.0
    Samaritan           0.0
    Lebanese_Christian  0.0
    Lebanese_Muslim     0.0
    Druze               0.0
    Syrian              0.0
    Yoruba              0.0
    Bantu_SA_Herero     0.0
    Bantu_SA_Ovambo     0.0
    Bantu_SA_Pedi       0.0
    Bantu_SA_S_Sotho    0.0
    Bantu_SA_Zulu       0.0
    
    77% Copt, 16% Southern-Levantine and 7% West-South African. 
    
    This first model was pretty crazy... I decided to include every Levantine group out there (xJews & Cypriots) as well as some West-African samples and every Bantu speaker sample available and I expected to get some overfit and wonky result (5% Druze, 6% Palestinian, 2% Lebanese and a fit like 0.03... or something) but the distance is not in overfitting territory and the model surprisingly only chose Palestinians among the Levantines. Seems sensible to me, to be honest.
    
    --
    
    [1] "distance%=0.0354 / distance=0.000354"
    
             Egyptian
                            
    Yemenite-Jew       40.90
    Moroccan           18.50
    Lebanese_Muslim    16.60
    Syrian             10.70
    Libyan              4.95
    Yoruba              2.80
    Saudi               2.10
    Mende_Sierra_Leone  1.25
    Bantu_SA_Ovambo     1.05
    Bantu_SA_Pedi       1.05
    Arab_Israel_2       0.10
    
    43% Arabian, 27% Northern-Levantine, 24% Maghrebi and 6% West-South African.
    
    Decided to include Arabians like the Saudi and Yemenite Jewish samples as well as some  Maghrebis here and see how they'd change things up and, as you can see, the results are interesting but I'd say far-less sensible now. The Arabians (YJs mainly) seem to have usurped Copts and there's way too much Levantine (Northern-Levantine, no less) and Maghrebi to make sense. The distance wreaks of overfitting too.
    
    --
    
    [1] "distance%=0.0968 / distance=0.000968"
    
             Egyptian
                            
    Syrian             25.40
    Samaritan          18.85
    Negev-Bedouin-B    16.35
    Moroccan           14.45
    Libyan             13.80
    Egyptian_Copt       5.20
    Bantu_SA_Ovambo     4.95
    Yoruba              1.00
    
    61% Levantine, 28% Maghrebi, 5% Copt and 6% West-South African.
    
    Believe all I did here was add Bedouin-Bs and remove YJs and Saudis. As nonsensical to me as the last result.
    
    --
    
    [1] "distance%=0.4552 / distance=0.004552"
    
             Egyptian
                           
    Libyan             56.1
    Palestinian        21.9
    Syrian             20.9
    Egyptian_Copt       1.2
    
    56% Maghrebi, 43% Levantine and 1% Copt.
    
    Think I removed the Negevians here but I also definitely removed every Sub-Saharan and/or Sahelian population (I wanted to see, just for fun (not because it would make any sense), what'd happen if Maghrebis proved to be the only thing the models could turn to in order to explain these Egyptians' SSA&orSahelian ancestry. Libyans ended up muddying up everything. As some may know, especially along the east of the country, Libyans often prove to be the most West-Asian or Egyptian shifted Maghrebis.
    
    --
    
    [1] "distance%=0.752 / distance=0.00752"
    
             Egyptian
                            
    Palestinian        46.05
    Egyptian_Copt      30.60
    Moroccan           21.20
    Syrian              2.15
    
    46% Southern-Levantine, 31% Copt, 21% Maghrebi and 2% Northern-Levantine.
    
    I removed Libyans here and the run became more interesting, I guess. But eh, I was just having fun at this point... Don't take this result or almost any after the first one seriously at all.
    
    --
    
    [1] "distance%=0.7781 / distance=0.007781"
    
             Egyptian
                           
    Egyptian_Copt      46.9
    Syrian             33.2
    Moroccan           19.9
    
    47% Copt, 33% Northern-Levantine and 20% Maghrebi. 
    
    Think all I did here was remove Palestinians.
    
    --
    
    [1] "distance%=0.525 / distance=0.00525"
    
             Egyptian
                           
    Egyptian_Copt      77.1
    Samaritan          13.4
    Mende_Sierra_Leone  4.6
    Bantu_SA_Tswana     2.8
    Syrian              2.1
    
    77% Copt, 16% Levantine and 7% West-South African.
    
    And finally, I wanted to see what'd happen now if I re-added the West African and Bantu-speaking groups and got another result I think is somewhat sensible. 6-7% West-African related ancestry (most likely owed to the slave-trade?) and ~15% Levantine (mainly Soutern-Levantine again) related admixture.
    I'm not all that familiar at all with the Y-DNA and mtDNA differences between Copts and other Egyptian so I'll have to leave it to other folks to comment on whether or not ~15% Southern-Levantine-like admixture really makes sense. Anyway, feel free to share some models you made using nee4speed111/Egyptian_Copt and make some suggestions models for me to run (which samples you think I should put in the input and target folders and so on) and I'll try to run 'em when I'm free, though I suggest just figuring out nMonte for yourself as well (it's easy to run).
    Hey Awale, I've sent you the coptic genomes by email. What do you think is preventing any of the models from being good fits for me, Especially considering that muslim Egyptian samples can be fitted pretty well?

  14. The Following 2 Users Say Thank You to nee4speed111 For This Useful Post:

     Agamemnon (11-15-2017), Megalophias (11-15-2017)

  15. #18
    Registered Users
    Posts
    55
    Sex
    Omitted
    Y-DNA
    J2b2-L283>Z1296+
    mtDNA
    H2a1

    Do Copts ever show evidence of Dinka or other Nilo-Saharan ancestry via D stats?

  16. The Following User Says Thank You to TuaMan For This Useful Post:

     nee4speed111 (Yesterday)

  17. #19
    Registered Users
    Posts
    265
    Sex

    Egypt
    Quote Originally Posted by TuaMan View Post
    Do Copts ever show evidence of Dinka or other Nilo-Saharan ancestry via D stats?
    I would imagine we do, considering every other sort of model has copts showing it. I've never seen D stats of copts before though, so I could be wrong

Page 2 of 2 FirstFirst 12

Similar Threads

  1. Modeling Slavs as mixtures of >100 ancients
    By Tomenable in forum Ancient (aDNA)
    Replies: 95
    Last Post: 09-03-2017, 07:23 PM
  2. Replies: 2
    Last Post: 01-25-2017, 11:27 AM
  3. Preliminary modeling of ancients
    By Kale in forum Ancient (aDNA)
    Replies: 2
    Last Post: 08-07-2016, 07:46 PM
  4. Iranian qpAdm modeling
    By Kurd in forum Persian
    Replies: 10
    Last Post: 07-25-2016, 05:56 AM
  5. Modeling S Asians with qpAdm
    By Kurd in forum Hindi
    Replies: 2
    Last Post: 07-21-2016, 08:57 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
  •