PDA

View Full Version : 23andMe - Middle Eastern Admixture



Kopfjäger
11-17-2014, 05:06 PM
Hi everyone,

I have a question regarding an inherited component illustrated through 23andMe's Ancestry Composition tool. According to the tool, 1.1% of my genome can be attributed to Middle Eastern ancestry, with 1.0% assigned to the Middle Eastern tapestry, while 0.1% is labeled as Broadly Middle Eastern and North African. There is no doubt I inherit this component from my mother, who is assigned 2.6% Middle Eastern ancestry.

My question is, what is the most likely source of the Middle Eastern component in my (and mother's) genome? My maternal grandmother was 100% Italian, with her family hailing from Naples and Avellino. I also had a great-uncle's yDNA tested (Haplogroup J2), and found that his only match is to a surname thought to be of Sicilian origin (Crimi), so I may have ancestry from there as well.

I appreciate any ideas or comments!

John Doe
11-17-2014, 05:26 PM
Hi everyone,

I have a question regarding an inherited component illustrated through 23andMe's Ancestry Composition tool. According to the tool, 1.1% of my genome can be attributed to Middle Eastern ancestry, with 1.0% assigned to the Middle Eastern tapestry, while 0.1% is labeled as Broadly Middle Eastern and North African. There is no doubt I inherit this component from my mother, who is assigned 2.6% Middle Eastern ancestry.

My question is, what is the most likely source of the Middle Eastern component in my (and mother's) genome? My maternal grandmother was 100% Italian, with her family hailing from Naples and Avellino. I also had a great-uncle's yDNA tested (Haplogroup J2), and found that his only match is to a surname thought to be of Sicilian origin (Crimi), so I may have ancestry from there as well.

I appreciate any ideas or comments, and appreciate it!
I get 95.1% Ashkenazi on standard, and while it's all classified as European, I wouldn't be surprised if at least 50% or more of that would be middle eastern.

Kopfjäger
11-17-2014, 09:52 PM
I just realized that this thread should have been posted to the "23andMe" sub-forum.

Kopfjäger
11-18-2014, 01:41 AM
Hi everyone,

I have a question regarding an inherited component illustrated through 23andMe's Ancestry Composition tool. According to the tool, 1.1% of my genome can be attributed to Middle Eastern ancestry, with 1.0% assigned to the Middle Eastern tapestry, while 0.1% is labeled as Broadly Middle Eastern and North African. There is no doubt I inherit this component from my mother, who is assigned 2.6% Middle Eastern ancestry.

My question is, what is the most likely source of the Middle Eastern component in my (and mother's) genome? My maternal grandmother was 100% Italian, with her family hailing from Naples and Avellino. I also had a great-uncle's yDNA tested (Haplogroup J2), and found that his only match is to a surname thought to be of Sicilian origin (Crimi), so I may have ancestry from there as well.

I appreciate any ideas or comments!

Would this Middle Eastern component be reflective of much deeper ancestry, such as the Neolithic, or could it also be a signal of more recent admixture? I know Southern Italy has connections with Phoenicians and Arabs.

DMXX
11-18-2014, 04:01 AM
There are three approaches you could take to better define what this Near-Eastern signal represents. Using all of them together will serve as internal validation (i.e. confirming whether the answer for one is repeated through the other despite both being different approaches):

1) Run several chromosome painters on GEDmatch. Compare the region with the Near-Eastern signal in 23andMe (select "Chromosome View") with those painters to confirm whether or not it's consistently from the Near-East or it is actually South European in origin and "masked" by the reference populations (please note 23andMe's Near-Eastern reference is defined by Iran, Syria and Turkey, all of which tend to get a fair bit of "Mediterranean" component score through ADMIXTURE).

2) Choose an open-genome project calculator with an Oracle program. Obtain your component scores through it. Then, create a "synthetic" version of yourself with the proportions of your known ancestry (f.ex. for a half Swedish half Turkish from Istanbul individual, they'd need to obtain the results for each, half them and then combine the final scores). This would give you a theoretical version of yourself independent of any additional ancestries. Separately run both your results and your "synthetic" version through the calculator's specific Oracle (up to 100-150 populations) and compare the results. Specifically, you'd be looking for any significant (e.g. >2 GD) variations with South European and Near-Eastern populations.

3) More complicated, but IBD segment sharing in the approximate area in the relevant chromosome will give you the most definitive answer. I suppose contacting Davidski for his services would be an option here.

Kopfjäger
11-18-2014, 11:44 PM
There are three approaches you could take to better define what this Near-Eastern signal represents. Using all of them together will serve as internal validation (i.e. confirming whether the answer for one is repeated through the other despite both being different approaches):

1) Run several chromosome painters on GEDmatch. Compare the region with the Near-Eastern signal in 23andMe (select "Chromosome View") with those painters to confirm whether or not it's consistently from the Near-East or it is actually South European in origin and "masked" by the reference populations (please note 23andMe's Near-Eastern reference is defined by Iran, Syria and Turkey, all of which tend to get a fair bit of "Mediterranean" component score through ADMIXTURE).

2) Choose an open-genome project calculator with an Oracle program. Obtain your component scores through it. Then, create a "synthetic" version of yourself with the proportions of your known ancestry (f.ex. for a half Swedish half Turkish from Istanbul individual, they'd need to obtain the results for each, half them and then combine the final scores). This would give you a theoretical version of yourself independent of any additional ancestries. Separately run both your results and your "synthetic" version through the calculator's specific Oracle (up to 100-150 populations) and compare the results. Specifically, you'd be looking for any significant (e.g. >2 GD) variations with South European and Near-Eastern populations.

3) More complicated, but IBD segment sharing in the approximate area in the relevant chromosome will give you the most definitive answer. I suppose contacting Davidski for his services would be an option here.

Thank you for the reply, DMXX. How can I access GEDmatch?

dp
11-19-2014, 12:55 AM
go to GEDmatch.com

Register an account (it goes by email)
They accept autosomal zipped Build 36 files from Ancestry, 23andme and FamilyTreeDNA.
Admixture and one-to-one are available after a few minutes. One to many searches may take a day or two.
hope that helps. I only had a minute.
dp :-)