DIYDodecad 2.0

[BenTurin]

Thank you nathanm!

[QUOTE=nathanm;329716]There are four possible bases in DNA, abbreviated by A, C, G, and T. Because of their molecular structure, A pairs with T, and C with G.

Nathan thank you so much! I keep coming back to you post and learning more each time!

[GayeSherman]

Target practice...

I've been having lots of fun with Dodecad and Dodecad Oracle.

One thing I did was to get the admixture for a 30Mb segment on my son's Chr 4. Dr. McDonald has it as African. I was looking for some American as there's the story of my MIL's "Indian grandmother" to figure out. I'm now thinking along the lines of she was more African than Indian.

Dodecad gave me the following for that segment:


3.01 East_European
13.95 West_European
32.12 Mediterranean
23.77 Neo_African
1.07 West_Asian
- South_Asian
- Northeast_Asian
- Southeast_Asian
5.76 East_African
2.66 Southwest_Asian
- Northwest_African
17.65 Palaeo_African

Oracle mixedmode gave this:

[1,] 53.9% Sardinian + 46.1% Nguni 5.1515
[2,] 53.5% Sardinian + 46.5% Pedi 5.3423
[3,] 54.6% Sardinian + 45.4% Fang 5.5747
[4,] 54.7% Sardinian + 45.3% Kongo 5.7390
[5,] 46.4% LWK + 53.6% Sardinian 5.8483
[6,] 54.7% Sardinian + 45.3% Xhosa 5.9330
[7,] 45% Bamoun + 55% Sardinian 6.1089
[8,] 53.1% Sardinian + 46.9% Luhya_X 6.2266
[9,] 53.3% Sardinian + 46.7% Luhya 6.3236
[10,] 54.4% Sardinian + 45.6% Kaba 6.3836
[11,] 53.7% Sardinian + 46.3% Sotho_Tswana 6.6542
[12,] 55.4% Sardinian + 44.6% Igbo 7.0552
[13,] 55% Sardinian + 45% Hausa 7.0708
[14,] 55.4% Sardinian + 44.6% Brong 7.1530
[15,] 52.8% Sardinian + 47.2% Alur 7.5137
[16,] 53.6% Sardinian + 46.4% Mada 7.8474
[17,] 56.1% Sardinian + 43.9% Bambaran 7.9826
[18,] 54.7% North_Italian + 45.3% Nguni 8.6002
[19,] 54.3% North_Italian + 45.7% Pedi 8.6324
[20,] 55.4% North_Italian + 44.6% Fang 8.7845
[21,] 56.6% Sardinian + 43.4% Dogon 8.8115
[22,] 55.5% North_Italian + 44.5% Kongo 8.8943
[23,] 55.5% North_Italian + 44.5% Xhosa 9.0074
[24,] 45.6% LWK + 54.4% North_Italian 9.0564
[25,] 52.7% ASW + 47.3% Sardinian 9.1128
[26,] 44.2% Bamoun + 55.8% North_Italian 9.1329
[27,] 57.2% Sardinian + 42.8% Yoruba_H 9.1792
[28,] 53.8% North_Italian + 46.2% Luhya_X 9.3278
[29,] 55.1% North_Italian + 44.9% Kaba 9.3611
[30,] 54.1% North_Italian + 45.9% Luhya 9.3959


Here's my question: Since I'm 50% Ashkenazi and 50% British/German (colonial American) with no NA or African admixture - where is MY contribution to this piece of his DNA? Is there any way I could tease it out? Am I the "Sardinian" piece?

My admixture for that section is:

East_European 23.79
West_European 14.36
Mediterranean 41.62
Neo_African -
West_Asian 2.82
South_Asian -
Northeast_Asian 0.34
Southeast_Asian -
East_African 2.30
Southwest_Asian 14.67
Northwest_African -
Palaeo_African -


His father is 50% Southern Italian, 25% German and the rest is Irish and the alleged "Indian" great-grandmother.


Thanks!

Gaye

[BenTurin]

Hi Gaye - how do you run Dodecad and Dodecad Oracle for one segment? I have one segment in common with a bunch of my matches and would like to do the same. Thanks!

I've been having lots of fun with Dodecad and Dodecad Oracle.

One thing I did was to get the admixture for a 30Mb segment on my son's Chr 4. Dr. McDonald has it as African. I was looking for some American as there's the story of my MIL's "Indian grandmother" to figure out. I'm now thinking along the lines of she was more African than Indian.

Dodecad gave me the following for that segment:


3.01 East_European
13.95 West_European
32.12 Mediterranean
23.77 Neo_African
1.07 West_Asian
- South_Asian
- Northeast_Asian
- Southeast_Asian
5.76 East_African
2.66 Southwest_Asian
- Northwest_African
17.65 Palaeo_African

Oracle mixedmode gave this:

[1,] 53.9% Sardinian + 46.1% Nguni 5.1515
[2,] 53.5% Sardinian + 46.5% Pedi 5.3423
[3,] 54.6% Sardinian + 45.4% Fang 5.5747
[4,] 54.7% Sardinian + 45.3% Kongo 5.7390
[5,] 46.4% LWK + 53.6% Sardinian 5.8483
[6,] 54.7% Sardinian + 45.3% Xhosa 5.9330
[7,] 45% Bamoun + 55% Sardinian 6.1089
[8,] 53.1% Sardinian + 46.9% Luhya_X 6.2266
[9,] 53.3% Sardinian + 46.7% Luhya 6.3236
[10,] 54.4% Sardinian + 45.6% Kaba 6.3836
[11,] 53.7% Sardinian + 46.3% Sotho_Tswana 6.6542
[12,] 55.4% Sardinian + 44.6% Igbo 7.0552
[13,] 55% Sardinian + 45% Hausa 7.0708
[14,] 55.4% Sardinian + 44.6% Brong 7.1530
[15,] 52.8% Sardinian + 47.2% Alur 7.5137
[16,] 53.6% Sardinian + 46.4% Mada 7.8474
[17,] 56.1% Sardinian + 43.9% Bambaran 7.9826
[18,] 54.7% North_Italian + 45.3% Nguni 8.6002
[19,] 54.3% North_Italian + 45.7% Pedi 8.6324
[20,] 55.4% North_Italian + 44.6% Fang 8.7845
[21,] 56.6% Sardinian + 43.4% Dogon 8.8115
[22,] 55.5% North_Italian + 44.5% Kongo 8.8943
[23,] 55.5% North_Italian + 44.5% Xhosa 9.0074
[24,] 45.6% LWK + 54.4% North_Italian 9.0564
[25,] 52.7% ASW + 47.3% Sardinian 9.1128
[26,] 44.2% Bamoun + 55.8% North_Italian 9.1329
[27,] 57.2% Sardinian + 42.8% Yoruba_H 9.1792
[28,] 53.8% North_Italian + 46.2% Luhya_X 9.3278
[29,] 55.1% North_Italian + 44.9% Kaba 9.3611
[30,] 54.1% North_Italian + 45.9% Luhya 9.3959


Here's my question: Since I'm 50% Ashkenazi and 50% British/German (colonial American) with no NA or African admixture - where is MY contribution to this piece of his DNA? Is there any way I could tease it out? Am I the "Sardinian" piece?

My admixture for that section is:

East_European 23.79
West_European 14.36
Mediterranean 41.62
Neo_African -
West_Asian 2.82
South_Asian -
Northeast_Asian 0.34
Southeast_Asian -
East_African 2.30
Southwest_Asian 14.67
Northwest_African -
Palaeo_African -


His father is 50% Southern Italian, 25% German and the rest is Irish and the alleged "Indian" great-grandmother.


Thanks!

Gaye

[nathanm]

Hi Gaye - how do you run Dodecad and Dodecad Oracle for one segment? I have one segment in common with a bunch of my matches and would like to do the same. Thanks!

Look at my post at the top of page 3 of this thread. You'll need the start and end points for the segment, which you can get from FTDNA's chromosome browser. Just select the matches who share that segment, then click the link that says "Download to Excel." It gives you a spreadsheet with all the matching segments, so you'll want to just pick out the major one, and use the lowest starting point and highest end point between your matches.

[Taz85]

Nathan, You seem to know alot about this. Is it true that my medi data might not be medi, but mideastern?

[Linds]

So I just tried the Dodecad Oracle program, it's so cool!

The basic run of my results against the v3 database didn't give me too much of interest since I'm such a mutt. My top two matches were Tuscan_X and Tuscan_H (14.18 and 14.48). Does anyone know what the difference is between those two populations? Also, what is TSI stand for?

Here's my top 20:

[1,] "Tuscan_X" "14.1782"
[2,] "Tuscan_H" "14.4791"
[3,] "TSI" "14.6122"
[4,] "Slovenian" "14.6722"
[5,] "Hungarians" "18.2044"
[6,] "N_Italian_D" "19.9561"
[7,] "Romanians_14" "20.0138"
[8,] "O_Italian_D" "20.7527"
[9,] "CEU" "20.7652"
[10,] "N._European" "20.862"
[11,] "German_D" "21.2329"
[12,] "Ashkenazy_Jews" "21.3144"
[13,] "Balkans_D" "21.5779"
[14,] "French" "21.7322"
[15,] "French_D" "22.1585"
[16,] "Argyll_1KG" "22.2916"
[17,] "Orcadian" "22.9737"
[18,] "Ashkenazi_D" "23.0478"
[19,] "Orkney_1KG" "23.3281"
[20,] "Tuscan" "25.0015"


I really like the Mixed Mode part, very interesting results. None super close to reality (McDonald predicted 67% English, 32% Armenian), but I suppose in the same ballpark.

Here's my top 10 mixed mode results:

[1,] "31.9% Druze + 68.1% German_D" "2.1976"
[2,] "32.1% Cypriots + 67.9% German_D" "3.0008"
[3,] "49.9% Ashkenazy_Jews + 50.1% German_D" "3.0439"
[4,] "47.9% Ashkenazi_D + 52.1% German_D" "3.181"
[5,] "68.8% German_D + 31.2% Iraq_Jews" "3.3529"
[6,] "32.3% Azerbaijan_Jews + 67.7% German_D" "3.4434"
[7,] "44% Cypriots + 56% Swedish_D" "3.6221"
[8,] "68.2% German_D + 31.8% Iranian_Jews" "3.633"
[9,] "64.7% German_D + 35.3% Turkish_D" "3.9599"
[10,] "68.7% German_D + 31.3% Samaritians" "4.0122"

[nathanm]

Nathan, You seem to know alot about this. Is it true that my medi data might not be medi, but mideastern?

I don't understand the science behind it nearly as well as lots of people on this forum. Much of the genetics I've only learned since I got into this hobby last year. However, I do understand his "experiment" design, and now have a good working knowledge of how the software/algorithm behaves. But let's set all that aside and consider history for a moment.

Mediterranean and Middle Eastern (at least the western Levant) are fairly close, geographically, genetically, and otherwise. Remember, the Romans used to call the Mediterranean Mare Nostrum (Lat. Our Sea). It wasn't a barrier for them, but a means of transportation between their provinces surrounding the sea. The Roman Empire was rather cosmopolitan for its time. Lots of people from the Middle East traveled to the capital, and many ended up settling there. Some of the emperors even had Middle Eastern backgrounds.

And long before Rome was anything but an insignificant village, the Phoenicians, and later the Greeks, established colonies all along the Mediterranean coast. Then after the empire "fell," Muslims ruled Sicily and parts of southern Italy for centuries during the Middle Ages. When the Jews were expelled from Spain in 1492, many of them settled in Italy.

I could go on, but I think you get the point. There's been significant admixture in Italy from the Middle East, North Africa, and everywhere else there was a Roman presence. Your DNA could be reflecting influence from any one of those inputs, or more than one.

[nathanm]

So I just tried the Dodecad Oracle program, it's so cool!

The basic run of my results against the v3 database didn't give me too much of interest since I'm such a mutt. My top two matches were Tuscan_X and Tuscan_H (14.18 and 14.48). Does anyone know what the difference is between those two populations? Also, what is TSI stand for?

They're just two different reference populations from Tuscany, Italy. The datasets are associated with papers authored by Henn, et al., and the other by Xing, et al. TSI is yet another Tuscan reference population, from the International HapMap Project. There's a spreadsheet on Google Docs showing all the reference populations in the Dodecad project.

I really like the Mixed Mode part, very interesting results. None super close to reality (McDonald predicted 67% English, 32% Armenian), but I suppose in the same ballpark.

That's because Dr McDonald probably doesn't have any German reference populations, unless he's keeping the samples we've submitted to him and putting us into groups. English is probably the closest he's got to German, and the same reason most people with German backgrounds get Orcadian on PF. None of the big, academic datasets have any German samples.

But in the Dodecad project, any ethnic group with at least five participants (who actually meet his entry criteria) form new reference populations, denoted by "_D" at the end of the name. Since Dodecad has about 800 participants, he's been able to add 36 reference populations of his own. The "German_D" group has 15 members. And there's a separate "Mixed_Germanic_D," "Dutch_D," and individual populations for many of the surrounding countries, so he's able to better distinguish between intra-European populations.

[Brunetmj]

There's been significant admixture in Italy from the Middle East, North Africa, and everywhere else there was a Roman presence.

I never cared much for the Romans after they beat my home Team. The Gauls.

[Linds]

They're just two different reference populations from Tuscany, Italy. The datasets are associated with papers authored by Henn, et al., and the other by Xing, et al. TSI is yet another Tuscan reference population, from the International HapMap Project. There's a spreadsheet on Google Docs showing all the reference populations in the Dodecad project.

Thanks....I'm still perplexed by that - I guess because Tuscans are on the border between Northern/Western Europe and Southern Europe, and my own ancestry is a combination of those.

That's because Dr McDonald probably doesn't have any German reference populations, unless he's keeping the samples we've submitted to him and putting us into groups. English is probably the closest he's got to German, and the same reason most people with German backgrounds get Orcadian on PF. None of the big, academic datasets have any German samples.

I'm not German (well one paternal g-grandparent was 1/2 German...so I'd be 6.5% German??)), but rather English and Czech for my "Western European" parts (but Germany is in-between so I'm assuming that's why that is popping up).

Does Dr. McDonald have Czech population sets?? I'm assuming no based on my results and how he said "Czech comes out" in an English and Armenian mix. Czech and German are VERY close, and I noticed there's no Czech populations listed in Dodecad either, so perhaps the German is reading the Czech DNA.

[tkb]

My Results

mtDNA L3d

FTDNA PF
Western European - Orcadian 37.54% 0.00%
West African - Mandenka 62.46% 0.01%
DIY
----------------------------
FINAL ADMIXTURE PROPORTIONS:
----------------------------

3.95% East_European
19.67% West_European
6.72% Mediterranean
35.41% Neo_African
2.56% West_Asian
0.08% South_Asian
0.26% Northeast_Asian
1.96% Southeast_Asian
7.39% East_African
0.00% Southwest_Asian
1.76% Northwest_African
20.24% Palaeo_African

DodecadOracle

[1,] "ASW" "16.3526"
[2,] "Fulani" "26.3141"
[3,] "Siddi" "26.4813"
[4,] "Luhya_X" "30.1089"
[5,] "Luhya" "30.3167"
[6,] "LWK" "30.8097"
[7,] "Mada" "30.8486"
[8,] "Hema" "31.2069"
[9,] "Alur" "31.3631"
[10,] "Pedi" "31.6652"

DodecadOracle mixed mode

[1,] "77.9% ASW + 22.1% N._European" "2.882"
[2,] "78.3% ASW + 21.7% Argyll_1KG" "2.9836"
[3,] "78% ASW + 22% German_D" "3.0608"
[4,] "78.5% ASW + 21.5% Orkney_1KG" "3.1331"
[5,] "78.5% ASW + 21.5% Orcadian" "3.1423"
[6,] "78.2% ASW + 21.8% CEU" "3.1708"
[7,] "64.1% Kaba + 35.9% N._European" "3.3356"
[8,] "79.1% ASW + 20.9% Mixed_Germanic_D" "3.3744"
[9,] "35.4% Argyll_1KG + 64.6% Kaba" "3.3804"
[10,] "35.1% Orkney_1KG + 64.9% Kaba" "3.5379"

[tkb]

???

I have Excel 2010 anyone know how I can make a color chart? also when I input my matches where we share segments start - end do we share that full result? Thanks

here is one of my FF matches

Chromosome# 1 Position: 77830973 to 80475894
window size: 105 SNPs | 2.64Mb

0.00% East_European
0.00% West_European
40.22% Mediterranean
0.00% Neo_African
0.00% West_Asian
0.00% South_Asian
0.00% Northeast_Asian
0.00% Southeast_Asian
13.92% East_African
0.00% Southwest_Asian
45.86% Northwest_African
0.00% Palaeo_African
CPU time = 3.43 sec

Dodecad Oracle

[1,] "Irish_D" "55.5133"
[2,] "Norwegian_D" "55.8996"
[3,] "Swedish_D" "57.1792"
[4,] "Cornwall_1KG" "57.7238"
[5,] "British_Isles_D" "57.9678"
[6,] "British_D" "58.3273"
[7,] "Kent_1KG" "59.2748"
[8,] "Mixed_Germanic_D" "60.1689"
[9,] "Dutch_D" "60.7952"
[10,] "Argyll_1KG" "65.6232"

Dodecad Oracle mixed mode

[1,] "87.2% Irish_D + 12.8% Lezgins" "54.6697"
[2,] "14.1% Lezgins + 85.9% Norwegian_D" "54.8476"
[3,] "72.2% Irish_D + 27.8% Norwegian_D" "55.4461"
[4,] "3.5% Adygei + 96.5% Irish_D" "55.4468"
[5,] "Irish_D" "55.5133"
[6,] "FALSE" "55.5133"
[7,] "FALSE" "55.5133"
[8,] "FALSE" "55.5133"
[9,] "FALSE" "55.5133"
[10,] "FALSE" "55.5133"

DodecadOracle Irish_D mixed mode

[1,] "Irish_D" "0"
[2,] "79.3% Cornwall_1KG + 20.7% Norwegian_D" "2.563"
[3,] "87.5% Cornwall_1KG + 12.5% Swedish_D" "3.0457"
[4,] "82.3% British_D + 17.7% Norwegian_D" "3.1853"
[5,] "88.3% British_Isles_D + 11.7% Norwegian_D" "3.3098"
[6,] "54% British_Isles_D + 46% Cornwall_1KG" "3.4259"
[7,] "97.3% British_Isles_D + 2.7% Swedish_D" "3.5101"
[8,] "98.2% Cornwall_1KG + 1.8% FIN" "3.5149"
[9,] "British_Isles_D" "3.5285"
[10,] "FALSE" "3.5285"

[GayeSherman]

Personally, I would only do the target analysis for larger segments - at least 10cM.

I just compared my brother's byseg run line by line to the Dodecad populations. Even though he's 100% Ashkenazi, there were few individual segments that "matched" an Ashenazi reference population. I'm wondering if I should run a byseg of something greater than 500/50 and see if that smoothes it out.

[Taz85]

I'm running....

byseg
1000
100

It's been over an hour, and it's only on the 8th chromosome. I think it's going to take over 3 hrs. lol......Worth a shot. This will give me a better result right? Clearer results?

[Taz85]

It actually took around 1hr 28mins.

byseg
1000
100