PDA

View Full Version : Noneuropean Segments in Otherwise European People



greerpalmer
07-03-2017, 04:01 PM
Hi guys, I wanted to open up a discussion on the validity of "surprise" nonEuropean segments in your results.

Personally, livingDNA assigns me 4.1% South Asian Pashtun which is actually a fairly significant number (present in all modes Cautious-Complete). My paper ancestry is primarily English, German, Irish with French, Swiss, and Luxembourg Germanic influence as well as Scandinavian.

Other tests have placed me varying degrees of trace nonEuropean-- MyHeritage 2.5% North African, 23andMe .1% North African/Middle Eastern, FamilyTreeDNA 1% East Central Africa (formerly 5% Middle Eastern), but all could be dismissed as noise or admixture from the Southern European-Iberian, Balkan or Italian that consistently shows as a minority share of my admixture.

Pashtun is interesting because A. It's so specific B. The confidence appears high and C. its not commonly associated with Southern Europeans. Interestingly though, in Cautious mode Pashtun related ancestry includes Armenia and Cyprus, which is much more believable.

Thoughts? I've seen several predominately English people with Near East, Central Asian or African parts.

CuriousAboutStuff
07-03-2017, 04:12 PM
Hi guys, I wanted to open up a discussion on the validity of "surprise" nonEuropean segments in your results.

Personally, livingDNA assigns me 4.1% South Asian Pashtun which is actually a fairly significant number (present in all modes Cautious-Complete). My paper ancestry is primarily English, German, Irish with French, Swiss, and Luxembourg Germanic influence as well as Scandinavian.

Other tests have placed me varying degrees of trace nonEuropean-- MyHeritage 2.5% North African, 23andMe .1% North African/Middle Eastern, FamilyTreeDNA 1% East Central Africa (formerly 5% Middle Eastern), but all could be dismissed as noise or admixture from the Southern European-Iberian, Balkan or Italian that consistently shows as a minority share of my admixture.

Pashtun is interesting because A. It's so specific B. The confidence appears high and C. its not commonly associated with Southern Europeans. Interestingly though, in Cautious mode Pashtun related ancestry includes Armenia and Cyprus, which is much more believable.

Thoughts? I've seen several predominately English people with Near East, Central Asian or African parts.

How far have you gotten back with your paper ancestry?

I know it's possibly a long shot, but from a quick search, it says Pashtuns are primarily located in Pakistan. There has been military activity in Pakistan, including by the British, even in the World War years. Do you think it's a possibility that something may have gotten mixed in at that time?

07-03-2017, 04:29 PM
Pashtuns Are also in Afghanistan, all around the Kyber Pass area, I think it’s possible it could be a British connection, I know I personally have family stories about the Napoleonic wars, and bring a French woman back to the UK.
For you a few generations later moving to the States, sounds plausible, certainly if this keeps popping up and many tests.
I think even Prince William and co, have Indian dna in their tree and in paper trail, so why not.

greerpalmer
07-03-2017, 05:03 PM
How far have you gotten back with your paper ancestry?

I know it's possibly a long shot, but from a quick search, it says Pashtuns are primarily located in Pakistan. There has been military activity in Pakistan, including by the British, even in the World War years. Do you think it's a possibility that something may have gotten mixed in at that time?

I'm mostly colonial american with most of my ancestry coming over in the first half of the 1600's so recent War activity is not possible. I do have more recent Irish and German lines that I have trouble dating past the 1800's. I also have a great grandfather that came over from Cornwall, the most recent immigration in my tree, but most of his lines are fairly well documented.

I feel like I have a really good paper trail, but it only takes one line of unknown...

FionnSneachta
07-03-2017, 05:09 PM
Considering non-European DNA is appearing in every test that you've taken, I'd say that it is probably real but from a long time ago so that it would be unknown to your family today. Europe is connected to Asia and Africa so it is possible that someone from those areas got in the mix at some stage.

CuriousAboutStuff
07-03-2017, 05:24 PM
I'm mostly colonial american with most of my ancestry coming over in the first half of the 1600's so recent War activity is not possible. I do have more recent Irish and German lines that I have trouble dating past the 1800's. I also have a great grandfather that came over from Cornwall, the most recent immigration in my tree, but most of his lines are fairly well documented.

I feel like I have a really good paper trail, but it only takes one line of unknown...

Ah, I understand.

greerpalmer
07-03-2017, 05:31 PM
Anyone familiar with Romani people and how they may report on livingDNA? They're Indo-Aryan with origins northwest of India....I'd assume they were present in England, as well as Germany, France, etc.

Pylsteen
07-03-2017, 05:44 PM
I find 4% quite much. It may be colonial; it may be Romani, there were lots of such families in England; maybe professions can shed some light on it (stereotypically circus work, or artist things; I heard Charlie Chaplin had Roma ancestry, but I am not sure of it. Other professions may be broom maker or more ordinary things like shoe maker). I am not an expert on Roma though.

MacUalraig
07-03-2017, 05:50 PM
Pashtuns Are also in Afghanistan, all around the Kyber Pass area, I think it’s possible it could be a British connection, I know I personally have family stories about the Napoleonic wars, and bring a French woman back to the UK.
For you a few generations later moving to the States, sounds plausible, certainly if this keeps popping up and many tests.
I think even Prince William and co, have Indian dna in their tree and in paper trail, so why not.

And Seb Coe...

http://www.dailymail.co.uk/femail/article-1235341/Lord-Coe-Punjabi-Playboy.html

greerpalmer
07-03-2017, 05:52 PM
I find 4% quite much. It may be colonial; it may be Romani, there were lots of such families in England; maybe professions can shed some light on it.

Good thought, from what I've seen I come from heavy farmer, dairymen and laborer stock. I've never paid too much attention so I will going forward.

CannabisErectusHibernius
07-03-2017, 06:58 PM
I haven't done living dna, but most calculators, including the previous version of ftdna fam finder give me 2 to 4 percent Indian or South Asian. As far as I know, all my lines go back to county Mayo, Ireland. This segment on my chromosomes are shared with other Irish people. I'm not as adept as others at reading the genetic tea leaves, but I'm guessing it's an ancient connection, rather then colonial or Romani.

JohnHowellsTyrfro
07-03-2017, 09:07 PM
I've been chasing something similar for a long time, and see this question come up quite often, so I keep repeating myself. :)
As far as I know all my recent ancestry is rural British European. I consistently get North West Pakistan region percentages on various tests and calculators ( Kalash, Burusho, Baloch) as well as Steppe/Turkic. The possibilities that have come up are:-
a) Actual recent Asian ancestry, which seems very unlikely given ancestors mainly in the rural Welsh borders. That's the first thing people came up with "you must have an Indian ancestor". I don't believe it. :)
b) Roma - possible but as far as I know very difficult to confirm, although I'm no expert in such things. No-one has been able to say to me yes, you definitely carry Roma markers although I have asked various people who know more than me about such things. If I understand things correctly "Asian" DNA isn't even that evident amongst people with definite Roma ancestry but I'm happy to be corrected on that.
c) It's basically ancient, probably showing relatively high percentages of earlier ancestry. The remote populations I mentioned Kalash etc. apparently retain an "older" DNA signature if I understand things correctly. I also get Basque as well which is another relatively genetically isolated population. I've also read that some "Asian" it isn't that uncommon in European populations.
So my best guess is that it is ancient but why it shows through in a fairly significant way, I haven't a clue. John

LivingDNA complete, but some other tests and calculators are similar :-

Asia (Central) 1.4%
Northwest Caucasus 1.4%
Asia (South) 1.4%
Burusho 1.4%

greerpalmer
07-03-2017, 09:59 PM
Interesting. Yes I consistently get Baloch and Druze on gedmatch calculators, but I always dismissed until the commercial test gave me such a large quantity. Unfortunately I've only tested my parents on familytreeDNA which isn't the most reliable and our results make no sense side by side, but from what I've learned from their transfers is they are significantly more South-European, which makes me think perhaps I inherited an ancient piece of southern European that standing alone seems to originate further east.

Jessie
07-04-2017, 01:28 AM
I get 1.3 Pashtun and 1.4 Chechen. I don't think it is real at all. I've never had a hint of this on other tests. I also only have Ireland and Britain in my results and no other European region. I have all Irish ancestry with only Irish / and a couple of possible Scots names in my family tree also a possible Huguenot name. The only other thing I've got in a couple of tests was a small percentage of Ashkenazi and Native American so with all ancestry from Ireland I think the algorithms are assigning me something a bit more exotic dna instead of some other European.

sktibo
07-05-2017, 06:22 AM
1.4% Kurdish here. The romantic in me thinks it's a Jewish connection but the realist in me tells me it's a poor calculation and less than adequate reference populations

Coldmountains
07-05-2017, 07:57 AM
It is noise or pre-historical genetic affinty. Pashtuns get also some "British" on such calculators but any kind of direct British ancestry can be pretty much excluded. Europeans don't have any kind of Pashtun, Kalash or Central/South Asian ancestry. Romani are quite distinct from Pashtuns genetically but such calculators maybe misinterpret Romani ancestry as "pashtun"

Cinnamon orange
07-05-2017, 08:12 AM
If you do have colonial US ancestry then it is possible to have Roma or East Indian but not likely at the levels Living DNA gave you, unless from a very endogamous community, where it was able to be kept and reintroduced.
I have seen African Americans and European colonials show low levels on 23andme, most below one percent.

I would test at best 23andme or secondly Ancestry to see if your results are similar to FTDNA and Living DNA.

Kathlingram
07-05-2017, 11:57 AM
Interesting. Yes I consistently get Baloch and Druze on gedmatch calculators, but I always dismissed until the commercial test gave me such a large quantity. Unfortunately I've only tested my parents on familytreeDNA which isn't the most reliable and our results make no sense side by side, but from what I've learned from their transfers is they are significantly more South-European, which makes me think perhaps I inherited an ancient piece of southern European that standing alone seems to originate further east.

The more recent South European at FTDNA as evidenced recently is likely not accurate, most people feel..I show it and I have none for sure..I had none there when I first transferred in 2012 or so either..23andme is more accurate and then AncestryDNA at least as ISOGG reports it.. as someone said "less than adequate population samples"

Amerijoe
07-05-2017, 02:49 PM
b) Roma - possible but as far as I know very difficult to confirm, although I'm no expert in such things. No-one has been able to say to me yes, you definitely carry Roma markers although I have asked various people who know more than me about such things. If I understand things correctly "Asian" DNA isn't even that evident amongst people with definite Roma ancestry but I'm happy to be corrected on that.

John, this is out of an article in 'Scientific American', thought it may be of interest.

'But the Romani did not always keep to themselves. As they moved through Europe and set up settlements, they invariably met—and paired off with—local Europeans. And some groups, such as the Welsh Romani, show a relatively high rate of bringing locals—and their genetics—into their families.' :) Joe

greerpalmer
07-05-2017, 03:00 PM
The more recent South European at FTDNA as evidenced recently is likely not accurate, most people feel..I show it and I have none for sure..I had none there when I first transferred in 2012 or so either..23andme is more accurate and then AncestryDNA at least as ISOGG reports it.. as someone said "less than adequate population samples"

23andme, FamilyTreeDNA, DNA.land and livingDNA all assign me 1-5% South European compared to my inferred results from my parents of 12-21%--but I also know my inherited British Isles is overindexed which would account for a good chunk of uninherited southern European. 23andme, MyHeritage, FamilyTreeDNA and livingDNA asignm me .1-4.1% Asian and African which I believe may also tie-in.

Kathlingram
07-05-2017, 03:16 PM
The more recent South European at FTDNA as evidenced recently is likely not accurate, most people feel..I show it and I have none for sure..I had none there when I first transferred in 2012 or so either..23andme is more accurate and then AncestryDNA at least as ISOGG reports it.. as someone said "less than adequate population samples"

I did hear back from my upgraded question:"We are only able to establish the percentage similarity of your DNA to the population samples available in our reference dataset. Currently, in the family ancestry part of our testing package, we can compare with and report on your ancestry up to 80 worldwide regions.

Unfortunately, even our fine-scale algorithm cannot distinguish between some very genetically similar countries. In these cases, we group them into genetically-distinct clusters that often reflect natural geographical boundaries. In order to clearly distinguish results into more specific regions/countries, sample data fulfilling very specific criteria will have to be collected and included in our reference datasets. This is a process that takes years.

Many populations are themselves formed from the mixture of other populations. Our algorithm will attempt to match you directly to a population/region, where we have it in our reference database. If:
we do not have your region of interest and/or,
we have a low sample size for that region, or,
if there is further population structure within that population/region,
then in the interim, we are likely to assign some ancestry to the regions we do have that contribute to your region of interest."

Again I do appreciate their prompt responding to questions

JohnHowellsTyrfro
07-05-2017, 04:12 PM
John, this is out of an article in 'Scientific American', thought it may be of interest.

'But the Romani did not always keep to themselves. As they moved through Europe and set up settlements, they invariably met—and paired off with—local Europeans. And some groups, such as the Welsh Romani, show a relatively high rate of bringing locals—and their genetics—into their families.' :) Joe

Thanks Joe. I do have quite a bit of ancestry from Herefordshire which has a relatively high Romany/traveller population as well as the Welsh borders and Wales generally, so it is possible but no-one so far has come back with an answer to confirm it or otherwise and I have asked several people. It would be nice to know one way or another. John

Amerijoe
07-05-2017, 04:28 PM
Thanks Joe. I do have quite a bit of ancestry from Herefordshire which has a relatively high Romany/traveller population as well as the Welsh borders and Wales generally, so it is possible but no-one so far has come back with an answer to confirm it or otherwise and I have asked several people. It would be nice to know one way or another. John

Knowing one way or the other, I can absolutely relate to. You never know what's in the woodpile until you start digging. Some admixtures have been pushing me towards Northern Ireland, some towards Wales and now England. Can they all be right? Yes, to varying degrees, there is a lot of wood in that pile and some from various forests. Good hunting. :) Joe

greerpalmer
07-13-2017, 07:04 PM
LivingDNA complete, but some other tests and calculators are similar :-

Asia (Central) 1.4%
Northwest Caucasus 1.4%
Asia (South) 1.4%
Burusho 1.4%

How do these values and estimated populations change on Cautious? Are they combined into a more general population or are they unassigned?

greerpalmer
11-15-2017, 08:04 PM
Hello, Some of you may have noticed matches with Humza Dasti's kits. He recently e-mailed a many people who have reached out regarding matches to the kits.

"
· All HGDP samples are from Stanford University’s Human Genome Diversity Project. The samples are anonymous. Their background is unknown beyond that they are of a certain ethnic group. This includes the BALOCH, BRAHUI, KALASH, PATHAN, UYGHUR, HAZARA, SINDHI, and MAKRANI samples.
· All HG samples are from the International Genome Sample Resource’s 1000 Genomes project: http://www.internationalgenome.org/. These samples are also anonymous, however, some of them have data on specific location and pedigree. This includes the PUNJABI LAHORE and BEB (Bengali in Bangladesh) samples.
· The IRANIAN samples and IRANIAN BANDARI samples are all from, 2016 Broushaki et al., Early Neolithic genomes from the eastern Fertile Crescent. I will be uploading more samples from Iran.
· The TAJIKISTANI, TAJIK YAGHNOBI, TAJIK RUSHAN, and TAJIK SHUGNAN are all from, 2015 Yunusbayev et. al., The Genetic Legacy of the Expansion of Turkic-Speaking Nomads across Eurasia.

The sample quality varies from study to study. The more SNP overlap, the better the comparison. All the HGDP samples have high-coverage. 400K+ SNPs. You can see this in your gedmatch match list under the OVERLAP section.

GEDMATCH is sorting your matches by TOTAL cM, regardless of the OVERLAP. This is why so many of my samples are floating to the top of your lists. This list of is of my 23andme sample; which has the most SNP overlap with my Uncle’s sample from 23andme @ 500K SNP. While my LivingDNA and FTDNA sample are at the top (only because they are me), you can see the overlap is much less @ 107K and 299K, overlap respectively. Comparing matches when they have such a difference is an exercise in comparing APPLES to ORANGES, not APPLES to APPLES.

There are multiple issues:

· If you have a LivingDNA, Genecove, older 23andme sample (V1-V3) tested sample, you already have poor overlap with most of the samples.
· The IRANIAN samples all have poor overlap @ 100K SNPs.
· The TAJIK and PUNJABI samples are slightly better @ 200K SNPs.
· The HGDP samples are all quite good at @400K SNPs"

Thought this may add some commentary.

Thanks