23andme v4 vs v5 ancestry determination

[cps]

So I'm tested on v4 and my mom is on the v5 test. So per the delay at 23andme I'm not phased yet with my mom. My paper ancestry is 100% Greek so it is not easy to tell what I got from which parent.

But I did see something that just makes me question how well this whole thing works. I got ~95% European and another 4.5% middle eastern. Not atypical results for a Greek. My mom got ~90% European and 10% middle eastern. That normally wouldn't seem weird (4.5 is about half of 10) .. except my mom is Mainland Greek and her parents were born in Greece in the 1890's before the population exchange with Turkey. Plus my dad is Aegean Islander--- I should be getting the middle eastern from him NOT her.

I dig further. For me I look at chromosome 10. On both chromosomes the whole thing painted Southern European. One is 100% Balkan and the other is about 75% Balkan and the rest is broadly Southern European.

Now my mom... on Chromosome 10 both chromosomes are almost 100% painted Middle Eastern... there is a tiny (I mean tiny sliver on the ends of one that is painted broadly European -- but these slivers are about 2% of the whole chromosome).

Now this just doesn't make sense. According to 23andme I share 49.9% of my DNA with my mom. According to 23andme we are half identical on Chromosome 10. Then how could our chromosome paintings be so different!

Comments?

[lukaszM]

I'm interested is it possible to upload V5 kit to DNA.LAND? Have you tried?

[cps]

I'll try that

[Adam A]

What testing companies don't (but should) tell us is that Southern European (especially Italian and Greek) DNA is extremely similar to middle eastern DNA. Just remember that borders don't always define race

[greerpalmer]

I observed the same with my v4 kit and my moms v5 kit. One of my chromosomes is 100% British Isle and the corresponding chromosome on my mother is 100% French/German. No wonder it's taking months to phase...must be hard when the puzzle pieces don't match.

[ianz91]

What testing companies don't (but should) tell us is that Southern European (especially Italian and Greek) DNA is extremely similar to middle eastern DNA. Just remember that borders don't always define race

No offense but that is complete BS, Southern Europeans and Middle Easterners are not genetically similar at all.

[greerpalmer]

No offense but that is complete BS, Southern Europeans and Middle Easterners are not genetically similar at all.

I wouldn't discredit that so fast. On Ancestry, 41% of people of native Southern European descent also have Middle Eastern descent, 47% have Caucasus. Depending on the service and reference population set they can easily be confused.

[cps]

But I think you all miss the point. The same testing company calls the same DNA Middle Eastern on the V5 and then Balkan on the V4. These companies advertise the BS as know your ancestry. People fall for this. I have a cousin (100% Greek) who is trying to figure out who his Italian ancestor is and from which side of his family he got it from because he got 15% Italian. I keep telling him... there probably is no Italian ancestor and he got the Italian from BOTH sides.

[greerpalmer]

On 23andMe both Balkan and Italian segments have a fairly low precision (88%) and one of the lower recalls as well (50% & 42%) meaning it is less likely to assign the segments unless it is very certain the segments fit those populations. MENA has a high precision (95%) and middle of the road recall (83%) meaning it can accurately pin point MENA segments as MENA, but also has a higher chance of identifying other segments as MENA as well.

[cps]

taking a closer look at her raw data. I found quite a bit of the snps were not determined or not genotyped, especially on chromosome 10. On the chromosome browser, her chromosome 10 had only 12 snps labeled. Mine had 88 snps genotyped. Hers seems like a very low number compared to mine. The vast majority said not-genotyped... rather than not determined. Isn't that too few snps to determine ancestry?