Announcement

Collapse
No announcement yet.

DIYDodecad 2.0

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Originally posted by nathanm View Post
    So your French, Orcadian, Bedouin, and Druze results from PF are consistent with the "Mediterranean" result from Dodecad. Like Matt said above, this is still pretty bleeding edge science. That's why people are submitting their data to Dr McDonald, Dodecad, and other projects. Their methods are different, and might provide different insights into our ancestry. I've learned a lot about how this stuff works in my own experiments with Plink and Admixture, by following this tutorial.
    Except that Druze, Bedouin do not cluster with Mediterranean on FTDNA PF. They cluster under ME.

    Southern European
    Italian
    Sardinian
    Tuscan

    These are Mediterranean, Right?
    Last edited by Taz85; 18 August 2011, 03:06 AM.

    Comment


    • #17
      Originally posted by Taz85 View Post
      Except that Druze, Bedouin do not cluster with Mediterranean on FTDNA PF. They cluster under ME.

      Southern European
      Italian
      Sardinian
      Tuscan

      These are Mediterranean, Right?
      You're talking about two different, tangentially related concepts. Geographically, all those populations live near the Mediterranean Sea. But that has little to do with the labels chosen by FTDNA for grouping reference populations, nor Dienekes for the Dodecad project's ancestral components. And it has absolutely nothing to do with "clustering," which is a method of statistical analysis.

      For the purposes of classification, FTDNA placed Druze and Bedouin under the Middle Eastern Continental Group, and Italian, Sardinian, and Tuscan under the Europe Continental Group and Southern European Subcontinental Group. They could've chosen to arrange them differently if they'd wanted, it's merely a matter of preference. They don't use the label Mediterranean for any group, but even if they did, it still wouldn't be the same as the "Mediterranean" ancestral component in the Dodecad project.

      Dienekes probably chose the label Mediterranean for that ancestral component because of the populations with a higher share of it. However, the label was completely arbitrary. The component itself was inferred by the algorithm, based on the number requested (K=12), the SNPs tested, and the mix of samples. Use a higher or lower number for K, a different mix of samples or SNPs, or a different algorithm or software, and the components inferred would be different.

      The results you received from PF, Dr McDonald, and Dodecad are presented differently, but they're all basically consistent, because they use different systems of classification, different methods of analysis, and different reference populations. If you want a fourth opinion, join the Eurogenes Genetic Ancestry Project. The results may look different yet, but will probably be consistent with all your prior results.

      Comment


      • #18
        More Explanation about DIYDodecad 2.0

        Dienekes has a new blog post explaining how to use results from DIYDodecad 2.0. The most interesting point is he discourages running Dodecad Oracle on chromosome or segment results. That's what I was hoping could help determine which ancestors I should focus my attention for some segments I share with a bunch of people. He does give some general advice for just those kind of situations, though. Just use the ancestral components themselves.

        I ran the target mode in DIYDodecad for several shared segments. The ancestral components are quite varied compared to my overall, or by chromosome, analysis. I share segment A with 5 matches, all overlapping by more than 7 cM. Segment B is shared with 3 matches, with at least 8 cM in common. The rest are shared with 2 matches each, with at least 8 cM in common. However, one of the matches in each segment C and F are my closest relatives--predicted 3rd cousins who still haven't responded. I share more than 22 cM with each of them on those segments. They should be the easiest to figure out how we're related, which should then help with the other matches who overlap there.

        Comment


        • #19
          Originally posted by nathanm View Post
          Dienekes has a new blog post explaining how to use results from DIYDodecad 2.0. The most interesting point is he discourages running Dodecad Oracle on chromosome or segment results. That's what I was hoping could help determine which ancestors I should focus my attention for some segments I share with a bunch of people. He does give some general advice for just those kind of situations, though. Just use the ancestral components themselves.

          I ran the target mode in DIYDodecad for several shared segments. The ancestral components are quite varied compared to my overall, or by chromosome, analysis. I share segment A with 5 matches, all overlapping by more than 7 cM. Segment B is shared with 3 matches, with at least 8 cM in common. The rest are shared with 2 matches each, with at least 8 cM in common. However, one of the matches in each segment C and F are my closest relatives--predicted 3rd cousins who still haven't responded. I share more than 22 cM with each of them on those segments. They should be the easiest to figure out how we're related, which should then help with the other matches who overlap there.

          How did you make the chart??

          Comment


          • #20
            Originally posted by Taz85 View Post
            DIY shows me as 26.10 Met. PF shows me as 0%. I have family trees going back to the late 1700's to Italy. I don't think anything like 26% is small. That's a major error on PF's part.
            I saw your earlier post on page one again. Had another thought. It is possible to have some Mediterranean show up as Middle Eastern or North African, in fact my Sicilian/Southern Italian does. The area was full of migration, trade, etc., causing a large mix of ancestry that these tests pick up.

            I noticed in your signature all the non DIY tests are reading you have Jewish. That could be your Mediterranean being read as such...

            So no I wouldn't agree PF is why off, just they way the test differences work.

            http://www.familytreedna.com/faq/ans...spx?id=22#1177

            Matt.
            Last edited by mkdexter; 20 August 2011, 06:01 AM.

            Comment


            • #21
              Originally posted by Taz85 View Post
              How did you make the chart??
              Excel. I just ran DIYDodecad in target mode for the segments I share with multiple matches. You can get the starting and ending numbers in a downloadable .csv file using the chromosome browser. Then to run in target mode, replace the last line in the dv3.par file with, e.g. segment A:
              Code:
              target
              19
              33246428
              40281760
              Then I just made a bar chart from the ancestral components for all the segments. The only problem is the colors assigned automatically by Excel start looking like they almost repeat after six colors. It's hard to tell the difference between, say West Asian and Northwest African. But I'm too lazy to manually assign distinguishable colors for each one. Here's another chart of my by chromosome results, which don't vary as much compared to the overall values:


              By the way, I forgot to mention before, DIYDodecad 2 has a huge speed improvement over version 1. The earlier version took almost 45 minutes to run. I re-ran myself in genomewide mode for comparison, and it took less than 5 minutes! But the ancestral component values were no more than 0.01 different than the previous run.

              Comment


              • #22
                Originally posted by mkdexter View Post
                I saw your earlier post on page one again. Had another thought. It is possible to have some Mediterranean show up as Middle Eastern or North African, in fact my Sicilian/Southern Italian does. The area was full of migration, trade, etc., causing a large mix of ancestry that these tests pick up.

                I noticed in your signature all the non DIY tests are reading you have Jewish. That could be your Mediterranean being read as such...

                So no I wouldn't agree PF is why off, just they way the test differences work.

                http://www.familytreedna.com/faq/ans...spx?id=22#1177

                Matt.
                DIY shows me at West Asian, which is the Middle East, over 17%. EuroDNACalc was done by Dodecad. Dodecad Oracle shows me having a lot of Jewish/Middle East admixture, So it's not confusing it with Mediterranean. My top admixture is Slovenian/Tuscan. Here is part of my admixture. DIY shows me as Jewish as well, so I don't think the confusion is there. FTDNA, Mcdonald, Dodecad Just use a different terminology.

                [9,] "17.9% Ashkenazi_D + 82.1% Slovenian" "3.2196"
                [11,] "18.8% Ashkenazy_Jews + 81.2% Slovenian" "3.2992"
                [17,] "14.5% Morocco_Jews + 85.5% Slovenian" "3.7181"
                [30,] "12.4% Sephardic_Jews + 87.6% Slovenian" "4.1158"
                [33,] "10.1% Druze + 89.9% Slovenian" "4.126"
                [38,] "9.1% Armenian_D + 90.9% Slovenian" "4.1733"
                [44,] "10.2% Azerbaijan_Jews + 89.8% Slovenian" "4.2416"
                [47,] "8.6% Armenians_16 + 91.4% Slovenian" "4.2964"
                [49,] "8.8% Assyrian_D + 91.2% Slovenian" "4.3512"
                [50,] "9.6% Iraq_Jews + 90.4% Slovenian" "4.3701"
                [51,] "9.7% Samaritians + 90.3% Slovenian" "4.3927"
                [52,] "11.3% Turkish_D + 88.7% Slovenian" "4.4116"
                [54,] "9.8% Iranian_Jews + 90.2% Slovenian" "4.4492"
                [55,] "10.2% Georgia_Jews + 89.8% Slovenian" "4.4616"
                [57,] "10.8% Uzbekistan_Jews + 89.2% Slovenian" "4.5473"
                [60,] "10.1% Turks + 89.9% Slovenian" "4.6459"
                [62,] "9.7% Palestinian + 90.3% Slovenian" "4.7457"
                [63,] "9.7% Lebanese + 90.3% Slovenian" "4.7665"
                [65,] "9.9% Kurd + 90.1% Slovenian" "4.8291"
                [67,] "9.9% Kurd_D + 90.1% Slovenian" "4.8848"
                [69,] "8.9% Jordanians_19 + 91.1% Slovenian" "5.0083"
                [71,] "8.9% Syrians + 91.1% Slovenian" "5.0268"
                [73,] "8.6% Iranian_D + 91.4% Slovenian" "5.0871"
                [74,] "8.6% Iranians + 91.4% Slovenian" "5.092"
                [81,] "73.3% German_D + 26.7% Turkish_D" "5.3202"
                [86,] "8.6% Egypt + 91.4% Slovenian" "5.4808"
                [87,] "7.2% Adygei + 92.8% Slovenian" "5.4924"
                [93,] "6.5% Lezgins + 93.5% Slovenian" "5.702"
                [95,] "6.6% Bedouin + 93.4% Slovenian" "5.8142"
                [97,] "6.8% Egyptans + 93.2% Slovenian" "5.9034"
                [99,] "6.8% Yemenese + 93.2% Slovenian" "5.9612"
                [100,] "36.8% Ashkenazy_Jews + 63.2% German_D" "6.0112"
                [102,] "7.2% Libya + 92.8% Slovenian" "6.0294"
                [103,] "75.4% German_D + 24.6% Turks" "6.0385"
                [106,] "6.2% Makrani + 93.8% Slovenian" "6.1148"
                [107,] "23.3% Druze + 76.7% German_D" "6.1676"
                [109,] "6.3% North_African_D + 93.7% Slovenian" "6.2173"
                [110,] "5.8% Brahui + 94.2% Slovenian" "6.2283"
                [111,] "6.8% Morocco_N + 93.2% Slovenian" "6.231"
                [112,] "6.6% Algeria + 93.4% Slovenian" "6.2558"
                [114,] "35.2% Ashkenazi_D + 64.8% German_D" "6.2797"
                [116,] "5.7% Balochi + 94.3% Slovenian" "6.2839"
                [120,] "72.5% German_D + 27.5% Sephardic_Jews" "6.3393"
                [121,] "5.2% TUNISIA + 94.8% Slovenian" "6.3821"
                [122,] "76% German_D + 24% Lebanese" "6.4424"
                [123,] "4.4% Kalash + 95.6% Slovenian" "6.4717"
                [125,] "4.6% Bnei_Menashe_Jews + 95.4% Slovenian" "6.5136"
                [126,] "3.2% Yemen_Jews + 96.8% Slovenian" "6.5406"
                [128,] "4.3% Moroccans + 95.7% Slovenian" "6.5705"
                [129,] "4.4% Sahara_OCC + 95.6% Slovenian" "6.5773"
                [131,] "2.9% Saudis + 97.1% Slovenian" "6.6348"
                [133,] "4.3% Pathan + 95.7% Slovenian" "6.6835"
                [134,] "4.1% Morocco_S + 95.9% Slovenian" "6.695"
                [135,] "77.3% German_D + 22.7% Iraq_Jews" "6.7015"
                [136,] "76.9% German_D + 23.1% Iranian_Jews" "6.7038"
                [137,] "21% Armenian_D + 79% German_D" "6.7053"
                [138,] "4.4% Uzbeks + 95.6% Slovenian" "6.7311"
                [139,] "23.4% Azerbaijan_Jews + 76.6% German_D" "6.7407"
                [140,] "77.1% German_D + 22.9% Samaritians" "6.7574"
                [141,] "3.3% Sindhi + 96.7% Slovenian" "6.7869"
                [142,] "3.7% Hazara + 96.3% Slovenian" "6.7974"
                [143,] "3.3% Burusho + 96.7% Slovenian" "6.8124"
                [144,] "3.2% Pakistani + 96.8% Slovenian" "6.8158"
                [145,] "96.8% Slovenian + 3.2% Kashmiri_Pandit" "6.8254"
                [146,] "52.2% Ashkenazi_D + 47.8% FIN" "6.8275"
                [148,] "3.1% Uygur + 96.9% Slovenian" "6.8685"
                [149,] "2.5% Cochin_Jews + 97.5% Slovenian" "6.8768"
                [150,] "74.7% German_D + 25.3% Uzbekistan_Jews" "6.8812"
                [153,] "24% Georgia_Jews + 76% German_D" "6.9053"
                [156,] "20.9% Assyrian_D + 79.1% German_D" "6.9141"
                [159,] "2.1% Fulani + 97.9% Slovenian" "6.9333"
                [160,] "20.3% Armenians_16 + 79.7% German_D" "6.9363"
                [161,] "98.2% Slovenian + 1.8% Velama" "6.9393"
                [163,] "2.3% Nepalese + 97.7% Slovenian" "6.943"
                [164,] "1.7% East_African_D + 98.3% Slovenian" "6.9432"
                [166,] "1.7% Ethiopian_Jews + 98.3% Slovenian" "6.9523"
                [167,] "1.7% Ethiopians + 98.3% Slovenian" "6.9538"
                [168,] "53.9% Ashkenazy_Jews + 46.1% FIN" "6.9544"
                [169,] "1.5% Mozabite_H + 98.5% Slovenian" "6.9583"
                [170,] "1.4% Mozabite + 98.6% Slovenian" "6.9625"
                [172,] "98.2% Slovenian + 1.8% Meghawal" "6.9753"
                [173,] "98.4% Slovenian + 1.6% TN_Brahmin" "6.9812"
                [174,] "1.5% AP_Brahmin + 98.5% Slovenian" "6.9943"
                [176,] "69.8% German_D + 30.2% Morocco_Jews" "6.997"
                [178,] "98.4% Slovenian + 1.6% Siddi" "7.0029"
                [180,] "23.6% Ashkenazy_Jews + 76.4% Hungarians" "7.0039"
                [181,] "98.7% Slovenian + 1.3% Naidu" "7.0048"
                [182,] "1.4% Indian_D + 98.6% Slovenian" "7.0077"
                [183,] "1.6% Kyrgyzstani + 98.4% Slovenian" "7.0097"
                [184,] "98.5% Slovenian + 1.5% Vaish" "7.0098"
                [186,] "1.3% SANDAWE + 98.7% Slovenian" "7.0115"
                [187,] "1.3% Maasai + 98.7% Slovenian" "7.0121"
                [188,] "75.7% German_D + 24.3% Kurd" "7.013"
                [189,] "1.2% MKK + 98.8% Slovenian" "7.0184"
                [190,] "1.2% Hema + 98.8% Slovenian" "7.0188"
                [191,] "98.8% Slovenian + 1.2% Tharu" "7.0207"
                [192,] "1.2% INS + 98.8% Slovenian" "7.0208"
                [193,] "98.9% Slovenian + 1.1% Vysya" "7.0209"
                [194,] "0.9% Khmer_Cambodian + 99.1% Slovenian" "7.0247"
                [195,] "22.3% Ashkenazi_D + 77.7% Hungarians" "7.0256"
                [196,] "98.9% Slovenian + 1.1% Hallaki" "7.0266"
                [197,] "99% Slovenian + 1% Kamsali" "7.027"
                [198,] "1.2% Mongol + 98.8% Slovenian" "7.0277"
                [199,] "1.1% Bulala + 98.9% Slovenian" "7.0295"
                [200,] "1.1% HADZA + 98.9% Slovenian" "7.0296"
                [201,] "98.8% Slovenian + 1.2% Srivastava" "7.0299"
                Last edited by Taz85; 20 August 2011, 06:16 PM.

                Comment


                • #23
                  Originally posted by nathanm View Post
                  Then I just made a bar chart from the ancestral components for all the segments. The only problem is the colors assigned automatically by Excel start looking like they almost repeat after six colors.
                  I can't figure out how to stack the % per chromosome in Excel. Ive only used it a few times. Every time I try it, it makes a bar for each %. If that makes sense.

                  Comment


                  • #24
                    Originally posted by Taz85 View Post
                    FTDNA, Mcdonald, Dodecad Just use a different terminology.
                    Exactly!

                    Originally posted by Taz85 View Post
                    I can't figure out how to stack the % per chromosome in Excel. Ive only used it a few times. Every time I try it, it makes a bar for each %. If that makes sense.
                    I know exactly what you mean. You need to use a different chart type. There's the normal column/bar graph with individual bars for each category (my version calls it "clustered columns"), a normal stacked graph where the length of each bar can vary, and a 100% stacked graph. The latter is the one you want. How you do that is slightly different, depending on which version of Excel you have, and what platform you're running it on. I'm running Excel 2008 on MacOS, and sometimes Excel 2007 on Windows XP.

                    Comment


                    • #25
                      Stacked Bar Chart

                      Attached is my results as a stacked bar chart. I did not like the default colors so I clicked on the Legend and the right clicked on them individually and changed the colors so some of them like Northeast Asian (which I made black) stand out a little better.
                      Attached Files

                      Comment


                      • #26
                        Here's mine....
                        Attached Files

                        Comment


                        • #27
                          Originally posted by Taz85 View Post
                          DIY shows me at West Asian, which is the Middle East, over 17%. EuroDNACalc was done by Dodecad. Dodecad Oracle shows me having a lot of Jewish/Middle East admixture, So it's not confusing it with Mediterranean.
                          Matt is entirely correct. All BGA analysis, whether PF, McDonald or Dodecad, has a very hard time distinguishing European Mediterranean (Italian, Greek) from Near Eastern-based populations, especially Jewish.

                          I've posted in another thread here about the Dodecad analysis of my and my sister's 23andMe results. We are full siblings, with the same parents. Dodecad is extremely accurate in its estimate of my admixture, telling me I'm 85-95% Sicilian and something to the east, Slavic, especially the Balkans. And yet it regards my sister as about 80% Sephardic and 20% northern European. We have no indication at all of even one line that is Jewish.

                          The lesson in looking at this analysis is not to take anything literally, but to look for clues, especially those that fit in with your known ancestry.
                          Last edited by MMaddi; 20 August 2011, 11:52 PM.

                          Comment


                          • #28
                            Originally posted by MMaddi View Post
                            Matt is entirely correct. All BGA analysis, whether PF, McDonald or Dodecad, has a very hard time distinguishing European Mediterranean (Italian, Greek) from Near Eastern-based populations, especially Jewish.

                            I've posted in another thread here about the Dodecad analysis of my and my sister's 23andMe results. We are full siblings, with the same parents. Dodecad is extremely accurate in its estimate of my admixture, telling me I'm 85-95% Sicilian and something to the east, Slavic, especially the Balkans. And yet it regards my sister as about 80% Sephardic and 20% northern European. We have no indication at all of even one line that is Jewish.

                            The lesson in looking at this analysis is not to take anything literally, but to look for clues, especially those that fit in with your known ancestry.
                            The initial question had nothing to do with anyone distinguishing between Med and near east populations. Nathan basically gave me the answer I was looking for between "Mediterranean" used by DIY and the terminology FTDNA uses.

                            Comment


                            • #29
                              Help pls

                              Ok I've dowload all programs I did everything the readme file said and i'm stuck I have the RGui up and running DIYDodecadWin dv3.par' had status 2 now I dont understand what more 2 do can someone help PLS.?

                              This is where i'm stuck

                              12 ancestral populations
                              166462 total SNPs
                              At line 142 of file DIYDodecad.f90 (Unit 50 "genotype.txt")
                              Traceback: not available, compile with -ftrace=frame or -ftrace=full
                              Fortran runtime error: End of file
                              Warning message:
                              running command 'DIYDodecadWin dv3.par' had status 2

                              where do i go from here pls help I'm not that good at this but i am good with the computer been trying the pass 2 hours

                              Comment


                              • #30
                                I am not sure but here is a guess

                                Did you change the directory of the command prompt to the working directory? aka cd c:\DIY
                                Also is there a space between DIYDodecadwin and the dv3.par
                                c:\DIY>DIYDodecadwin dv3.par
                                I hope this helps.
                                Last edited by ORE; 21 August 2011, 02:14 AM. Reason: to add cd c:\DIY

                                Comment

                                Working...
                                X