Page 162 of 182 FirstFirst ... 62112152160161162163164172 ... LastLast
Results 1,611 to 1,620 of 1816

Thread: Davidski's Global 25 nMonte results

  1. #1611
    Registered Users
    Posts
    968
    Sex
    Location
    Florida, USA.
    Ethnicity
    English, Scottish & Irish
    Nationality
    American
    Y-DNA
    I-A13243
    mtDNA
    H1e2

    England Scotland Ireland United States of America Vatican Germany Schleswig-Holstein
    I know I’m probably pushing it a bit here but it was kind of fun anyway. I took the two Roman Gladiators who I’m supposed to be the closest to and then I added in the two Italian and Hungarian Medieval samples that match my Haplogroup (I-M253) and ran them along with the following groups:


    Pen 0.001
    England_Anglo-Saxon +England_Roman--3DT16 +England_Roman--6DT18 +Germany_Medieval +Hungary_Medieval_Szolad--SZ45 +Italy_Medieval_Collegno--CL63 +Sweden_Viking_Age

    Custom:AGUser_JMcB = 1.4122 — 9.17, 30.83, 10.83, 16.67, 13.33, 8.33, 10.83

    Is this a valid way to go about trying a run?
    Last edited by JMcB; 09-14-2018 at 12:57 PM.
    Known Paper Trail: 45.3% English, 29.7% Scottish, 12.5% Irish, 6.25% German & 6.25% Italian. Or: 87.5% British Isles, 6.25% German & 6.25% Italian.
    LivingDNA: 88.1% British Isles (59.7% English, 27% Scottish & 1.3% Irish), 5.9% Europe South (Aegian 3.4%, Tuscany 1.3%, Sardinia 1.1%), 4.4% Europe NW (Scandinavia) & 1.6% Europe East, (Mordovia).
    FT Big Y: I1-Z140 branch I-F2642 >Y1966 >Y3649 >A13241 >Y3647 >A13248 (circa 900 AD) >A13242/YSEQ (circa 1030 AD) >A13243/YSEQ (circa 1550 AD).

  2. #1612
    Registered Users
    Posts
    624
    Sex
    Location
    Netherlands
    Ethnicity
    South-Dutch
    Nationality
    Dutch
    Y-DNA
    I2a2a1b2-CTS1977
    mtDNA
    H13a1a1

    Netherlands Belgium
    I did some exploration on the Longobard data in Global25. I had two questions:
    1. Can I separate the Northern and Southern Longobard samples?
    2. Can Global25 differentiate between Longobard_Medieval and Germany_Medieval?

    To answer the first question I extracted the Longobard and Germany_Medieval data and I ran a cluster analysis with mclust.
    Mclust found a very sharp separation of majority sample and a minority cluster. The plot is here:
    North_South.jpg

    The minority cluster consists of CL23,CL36,CL57,CL94,CL121,SZ27,SZ28,SZ32,SZ36,SZ37 ,SZ40,SZ43.
    Of these 12 minority samples I could not find CL36 in the paper. The other 11 according the all had a considerable TNI component (Tuscan Italy).
    The German_Medieval samples were all in the majority cluster.
    So yes, the Longobard Southern samples can clearly be discriminated in the Global25 data.

    My next question is whether the Northern Longobard samples can be discriminated from the Germany_Medieval samples.
    After dropping the Southern samples from my selection, I first ran mclust; it saw only one cluster.
    Next I ran a supervised test (Random Forests). The result was that the Northern Longobards were all imputed correctly, but the Germany_Medieval samples were only 23% imputed correctly (77% were imputed as Longobard).
    This is a well known problem in classification of imbalanced data, the minority samples are classified as majority. Here I could use 10 Germany_Medievals samples and 33 Longobard_North samples.
    So it is very hard to discriminate between Longobard_North and Germany_Medieval when using Global25.

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

     Просигој (09-14-2018),  Bollox79 (09-14-2018),  Dewsloth (09-14-2018),  Finn (09-14-2018),  JMcB (09-14-2018),  palamede (09-14-2018),  ph2ter (09-14-2018),  randwulf (09-14-2018)

  4. #1613
    Banned
    Posts
    6,337
    Sex
    Location
    Torun
    Ethnicity
    Central 75% + 25% Mazovia
    Nationality
    Pole
    Y-DNA
    R1a > M198 > YP 1337

    Poland
    Quote Originally Posted by Huijbregts View Post
    So it is very hard to discriminate between Longobard_North and Germany_Medieval when using Global25.
    Probably both have same sort of ancestry.

    BTW I'm waiting for Alemanni in G25. In study their ADMIXTURE components are ridiculous (like they were pure Lithuanians and /or Estonians, which is absurd, considering they weren't Eastern Germanics and comparing them to Baiuvarii and Longobards). Even western Slavs like Sorbs or Poles are more NW-Euro then them on this ridiculous ADMIXTURE plot.
    And what was completely disappointing practically all are R1b... Which is not very NE-Euro characteristic.
    Last edited by lukaszM; 09-14-2018 at 02:23 PM.

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

     Bollox79 (09-14-2018),  Capitalis (09-14-2018),  Finn (09-14-2018),  JMcB (09-14-2018),  palamede (09-14-2018),  Theconqueror (09-14-2018)

  6. #1614
    Registered Users
    Posts
    731
    Sex
    Location
    Lisbon, Portugal
    Ethnicity
    Western Meseta Iberian
    Nationality
    Portuguese
    Y-DNA
    E-Y31991 > PF4428
    mtDNA
    H20

    Portugal 1143 Portugal 1485 Portugal Order of Christ
    Quote Originally Posted by randwulf View Post
    I tried out the new Lombards and other new references. Some of the tribe had affinity to them. I preferred in the models the more recent references, but did use others.

    Me:

    [1] "distance%=0.5346"

    Beaker_Central_Europe,33.8
    Germany_Medieval_IACD,21.8
    Germany_Medieval,21.4
    Barcin_N,13.6
    Italy_Medieval_Collegno,8.4
    Xiong_Nu,1

    Dad:

    [1] "distance%=0.7058"

    Germany_Medieval_IACD,30
    Petrovka_MLBA,25.6
    Baden_LCA,21.4
    England_MBA,16.2
    Levant_BA_South,3.2
    Man_Bac_all,2.6
    South_Africa_2000BP,1

    Mom:

    [1] "distance%=0.6277"

    England_Roman,44.2
    England_LBA,14.8
    Mentese_N,13.6
    Germany_Medieval,12.6
    Germany_Medieval_ACD,9.4
    Avar_Hungary_Szolad,5.4

    Wife:

    [1] "distance%=0.4838"

    Italy_Medieval_Collegno,51.4
    Poprad,40.6
    Natufian,3.6
    Wales_CA_EBA,2.4
    Germany_Medieval,1.4
    Italy_Medieval_Collegno_o1,0.6

    FIL:

    [1] "distance%=0.9076"

    Italy_Medieval_Collegno,50
    England_Roman,21
    Germany_Medieval,15.2
    Beaker_Britain,10.8
    Natufian,2
    Oakaie1_all,1
    Aren't those models very overfitted?
    G25 Hidden Content and Hidden Content distances
    Hidden Content
    Hidden Content
    Hidden Content

    DEIBABOR
    IGO
    DEIBOBOR
    VISSAIEIGO
    BOR

  7. #1615
    Registered Users
    Posts
    429
    Sex
    Omitted
    Location
    Cardiff
    Ethnicity
    NW European
    Nationality
    Welsh
    Y-DNA
    R1b-U106 (Z306/S497)
    mtDNA
    H13a1a

    England Ireland Wales European Union
    Quote Originally Posted by Bollox79 View Post
    Yes certainly more than the other men tested so I figured that meant Baltic ancestry before they arrived in the Isles. I also noticed that the Niederstorzingen men have high shared drift (?) With Scandinavian/Lithuanians is this becoming the norm with these early Germanics ?
    I added Baltic_IA to my model and they didn't accept it, so not sure now.

    Attachment 25906
    My ancestry: 53% S Eng, 2% N Eng, 39% Ire, 6% S Wal
    LivingDNA: 53% S Eng, 7% E Anglia, 15% SW Scot, 9% S Wal, 8% Brit Isles, 4.5% Basque/Sardinian/Tuscan, 3.4% Chechen/Indian/Kurdish

  8. The Following 5 Users Say Thank You to Capitalis For This Useful Post:

     Bollox79 (09-14-2018),  Camulogène Rix (09-14-2018),  Finn (09-14-2018),  JMcB (09-14-2018),  palamede (09-14-2018)

  9. #1616
    Registered Users
    Posts
    429
    Sex
    Omitted
    Location
    Cardiff
    Ethnicity
    NW European
    Nationality
    Welsh
    Y-DNA
    R1b-U106 (Z306/S497)
    mtDNA
    H13a1a

    England Ireland Wales European Union
    Quote Originally Posted by Dewsloth View Post
    Dad got:


    Dewsloth_dad
    dist: 1.4473
    England_IA 50
    Hallstatt_Bylany 32.5
    Balkans_IA 14.17
    Basque_Spanish 3.33
    Nordic_IA 0
    What is your father's ancestry?
    My ancestry: 53% S Eng, 2% N Eng, 39% Ire, 6% S Wal
    LivingDNA: 53% S Eng, 7% E Anglia, 15% SW Scot, 9% S Wal, 8% Brit Isles, 4.5% Basque/Sardinian/Tuscan, 3.4% Chechen/Indian/Kurdish

  10. The Following User Says Thank You to Capitalis For This Useful Post:

     Dewsloth (09-14-2018)

  11. #1617
    Registered Users
    Posts
    429
    Sex
    Omitted
    Location
    Cardiff
    Ethnicity
    NW European
    Nationality
    Welsh
    Y-DNA
    R1b-U106 (Z306/S497)
    mtDNA
    H13a1a

    England Ireland Wales European Union
    Quote Originally Posted by Molfish View Post
    An AD Model

    Fit 1.4027
    England Roman 54.17
    Germany Medieval 16.67
    Italy Medieval Collegno 25.83
    Sweden Viking Age 3.33

    Mother

    Fit 1.3361
    England_Roman 38.33
    Germany_Medieval 19.17
    Italy_Medieval_Collegno 35.83
    Sweden_Viking_Age 6.67

    Replacing Collegno with Szolad

    Fit 1.4094
    England Roman 31.67
    Germany Medieval 11.67
    Hungary Medieval Szolad 52.5
    Sweden Viking Age 4.17

    Mother

    Fit 1.0181
    England_Roman 29.17
    Germany_Medieval 3.33
    Hungary_Medieval_Szolad 61.67
    Sweden_Viking_Age 5.83
    Germany_Medieval and Sweden_Viking_Age are hugely Nordic shifted.

    See David's PCA (same as mine but I think you'll trust his more):

    http://eurogenes.blogspot.com/2018/0...diversity.html
    Last edited by Capitalis; 09-14-2018 at 09:37 AM. Reason: Link added
    My ancestry: 53% S Eng, 2% N Eng, 39% Ire, 6% S Wal
    LivingDNA: 53% S Eng, 7% E Anglia, 15% SW Scot, 9% S Wal, 8% Brit Isles, 4.5% Basque/Sardinian/Tuscan, 3.4% Chechen/Indian/Kurdish

  12. #1618
    Registered Users
    Posts
    655
    Sex
    Location
    Pennsylvania, USA
    Ethnicity
    NW European/Celto-Germani
    Nationality
    American
    Y-DNA
    R1b-U106-DF98
    mtDNA
    T2b2bIrish/Scan

    Ireland Scotland Germany Canada Netherlands Northern Ireland
    Quote Originally Posted by Capitalis View Post
    I added Baltic_IA to my model and they didn't accept it, so not sure now.

    Attachment 25906
    Hmm still Nordic IA is very high... for some like Germany/Medieval and 6drif-3 etc...

    I'm trying to get an idea of where 6drif-3's male ancestors may have been from - because even though we was fairly "local" per his isotopes - we have a S4004+ sample with German then Dutch ancestry - and you can see S4004 in the block of SNPs I share with 6drif-3: S4004/FGC14818/FGC14823-FGC14816/FGC14817 - only some SNPs off of S4004... like I said there are two Swedes who are also positive for the SNPs that I list with 6drif-3 here and a couple more with me before we branch off etc... this is our little group at Big Tree: https://www.ytree.net/DisplayTree.php?blockID=3336
    Cheers!
    Last edited by Bollox79; 09-14-2018 at 10:52 AM.
    Y-DNA: 4th GGF Adam Weaver born 1785 in Pennsylvania - Sergeant in US 17th Infantry, War of 1812: R1b-U106-Z381-Z156-Z305/306/307-Z304-DF98-S1911-S1894/S1900-S4004/FGC14818/FGC14823-FGC14816/FGC14817. I share these SNPs w/ York Gladiator 6drif-3!

    mtDNA: 3rd GGM Bridget Dana b. 1843 Ireland - T2b2b - Ireland, Scandinavia and Hungary - T2b female warrior Grave Bj 581 near Birka, Sweden. Relative of King Bela III of Hungary (his Y-DNA and autosomal kinsman buried near him had mtDNA T2b2b1)!

  13. The Following User Says Thank You to Bollox79 For This Useful Post:

     JMcB (09-14-2018)

  14. #1619
    Registered Users
    Posts
    1,734
    Sex
    Location
    Pennsylvania
    Ethnicity
    West European
    Nationality
    USA
    Y-DNA
    R1b-U152-Z36+FGC6511
    mtDNA
    H11a2a

    United States of America Germany England France Scotland Ireland
    Quote Originally Posted by Ruderico View Post
    Aren't those models very overfitted?
    From a distance perspective, no they are not. I am using unscaled data and these distances are good ones for unscaled results (the ideal being about 0.5 to 1 %). I removed as many references from running with the full spreadsheet as I could without the distance jumping. Also, most of the time , the references are reasonable/normal regional picks for our ancestry, so not surprising, either. I am more concerned about my choices with relation to crossing eras than these being over fit. It is my thinking that the reference sets are incomplete enough that if the tools are wanting to make my model, for example, include a Copper Age Baden reference with my mostly Iron Age to Medieval model for my father whose ancestry comes primarily from that area, then I am guessing it needs it and the medieval references aren't complete enough to cover the variation from that area. So, I am just letting the data go where it wants to within bounds so that I can learn something from it.

  15. The Following 2 Users Say Thank You to randwulf For This Useful Post:

     palamede (09-14-2018),  Ruderico (09-14-2018)

  16. #1620
    Registered Users
    Posts
    624
    Sex
    Location
    Netherlands
    Ethnicity
    South-Dutch
    Nationality
    Dutch
    Y-DNA
    I2a2a1b2-CTS1977
    mtDNA
    H13a1a1

    Netherlands Belgium
    Quote Originally Posted by randwulf View Post
    From a distance perspective, no they are not.
    I agree with Ruderico that these models appear overfitted. Distance may be a warning, but it is not a guarantee that the model is not overfitted.
    The models show unexpected results, which resemble what nMonte produces without penalizing.
    Did you calculate these results with your own tool? That may be more overfitting than nMonte3.

  17. The Following User Says Thank You to Huijbregts For This Useful Post:

     Ruderico (09-14-2018)

Page 162 of 182 FirstFirst ... 62112152160161162163164172 ... LastLast

Similar Threads

  1. Eurogenes Global 10 PCA and updated nMonte (a.k.a. nMonte3)
    By randwulf in forum Autosomal (auDNA)
    Replies: 40
    Last Post: 05-25-2018, 01:09 PM
  2. Davidski's Basal-rick K7 + Global 10 Genetic Map Results
    By khanabadoshi in forum Autosomal (auDNA)
    Replies: 831
    Last Post: 12-02-2017, 08:41 PM
  3. Replies: 73
    Last Post: 11-25-2017, 05:31 PM
  4. Replies: 466
    Last Post: 09-26-2017, 11:57 AM
  5. nMonte results
    By firemonkey in forum Autosomal (auDNA)
    Replies: 0
    Last Post: 05-25-2017, 09:50 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
  •