PDA

View Full Version : South asian with over 75% DNA relatives non-south asian



Ace Mumtaz
07-07-2017, 10:37 PM
recently when checking my DNA results i noticed a large portion of my DNA relatives where not south Asian despite being originally from Kashmir in here you can see the results for yourself


https://you.23andme.com/reports/dna_family/print/

I had 99.1% south Asian ancestry and 0.8% broadly European from 23 and me. My Harrapa World results from Gedmatch showed me to be just over 20% European including Caucasian, north eastern European etc. whats interesting is that a lot of my relatives have some French & German ancestry and some gedmath calculators said that I was around 5% German. can anyone explain these results? I don't get why despite being 99.1% South Asian, so many of my DNA relatives are European. surely if I have so many European relatives, it shows that my European ancestry is more than 0.8%.

serena297
08-06-2017, 03:43 AM
I've had the same problem and honestly I think it's just old overlap in common Indo-European markers? I mean I don't know much about this, but that's how I rationalize it. I've had the same thing happen over on GEDmatch and I just chalk it up to shared old ancestry rather than newer ancestry? GEDmatch is amusing when it tries to figure out what the "other" is. If you know your family history, have a family tree perhaps, etc. I'd say that'd be your best bet to try to bounch things off of to get definite answers!

evon
08-06-2017, 09:19 AM
recently when checking my DNA results i noticed a large portion of my DNA relatives where not south Asian despite being originally from Kashmir in here you can see the results for yourself

I had 99.1% south Asian ancestry and 0.8% broadly European from 23 and me. My Harrapa World results from Gedmatch showed me to be just over 20% European including Caucasian, north eastern European etc. whats interesting is that a lot of my relatives have some French & German ancestry and some gedmath calculators said that I was around 5% German. can anyone explain these results? I don't get why despite being 99.1% South Asian, so many of my DNA relatives are European. surely if I have so many European relatives, it shows that my European ancestry is more than 0.8%.

The most obvious explanation is that Americans, Brits, Irish etc are over-represented in all DNA databases, thus "everybody" will get matches with these populations. There are also a lot of people in Asia that have some form of ancestry from colonial times, in the case of south Asia I would guess British, but could also be French, Danish etc... Lastly, Romani ancestry do sometimes mean that people get matches in both west and south Asia.. We have a few smaller south Asian matches, I have not figured out how we match (seems like a fruitless endeavor), but I suspect it is from colonial times...

kush
08-06-2017, 10:53 AM
White relatives could just be what serena297 said which is the shared indo-european markers, and there aren't many indians or pakistanis on 23andme. When you click oracles on harappa, or any other calculator you are using on gedmatch, you'll get population matches and distances, and most likely you will get kashmiri and punjabi as your closest population matches since most people for those regions get similar components as you. Including the euro+ caucasian percentages so you most likely have no british or euro blood from colonizers. The ne-euro in south asians is through steppe migrations from central asia 4-5 thousand years ago, not recently from colonizers.

Here's mine for example and I'm from andhra pradesh in south india. I get no NE-euro in my results since I lack the steppe component. can you post your results and oracles? thanks.

Admix Results (sorted):

# Population Percent
1 S-Indian 53.7
2 Baloch 37.19
3 Caucasian 4.32
4 NE-Asian 1.52
5 SE-Asian 1.2
6 SW-Asian 1.1
7 Papuan 0.45
8 American 0.4
9 Mediterranean 0.12

Single Population Sharing:

# Population (source) Distance
1 ap-reddy (harappa) 1.94
2 velama (metspalu) 2.97
3 karnataka (harappa) 3.42
4 velama (reich) 3.51
5 singapore-indian-b (sgvp) 4.91
6 andhra-pradesh (harappa) 4.97
7 gujarati-a (1000genomes) 5.19
8 tharu (metspalu) 5.36
9 tamil (harappa) 5.4
10 gujarati-patel (harappa) 6.01
11 ap-hyderabad (harappa) 6.07
12 gujarati-a (hapmap) 6.08
13 tamil-vellalar (harappa) 6.31
14 kerala-christian (harappa) 6.4
15 kerala-nair (harappa) 6.49
16 kerala-muslim (harappa) 6.5
17 sinhalese (harappa) 6.51
18 dharkar (metspalu) 6.53
19 up-muslim (metspalu) 6.67
20 rajasthani (harappa) 6.81

evon
08-06-2017, 11:30 AM
White relatives could just be what serena297 said which is the shared indo-european markers, and there aren't many indians or pakistanis on 23andme.

That is unlikely, shared segments that old would be tiny 1-3cM and have very few SNP, way to low for a match over at 23andme, and only visible at Gedmatch if one intentionally lowered the threshold below the default parameters..

wombatofthenorth
08-07-2017, 01:37 AM
It's testing bias. The database has mostly testers from the US and next most from Canada and the UK. And then from an area where you from it has very, very few testers. So if the database has 200+ people from outside your region for every 1 from within it, your results list will get radically skewed.