View Full Version : Ancestry Composition X Chromosome inconsistency between father and daughters

08-03-2016, 06:30 PM
Okay, here's a question:

Why would Ancestry Composition show inconsistent results between a father and his daughters on the X chromosome?

To be clear, the results aren't totally inconsistent. That is, my father and all three of his daughters show "Southern European" on the X chromosome, in the same region. This encompasses roughly the first three-fourths of the chromosome. The remainder shows up as "broadly Northwestern European", but nothing more specific.

However, drilling down a bit deeper, the bulk of the Southern European portion of my father's X chromosome shows as "Italian" -- probably about two-thirds of it, or around half of the chromosome. This segment accounts for all of my father's 1.3% Italian.

Meanwhile, none of my sisters shows any Italian on the X chromosome. My youngest sister, in fact, shows nothing more specific than "broadly Southern European" (in addition, of course, to the "broadly Northwestern European" part). The other two sisters also show a little bit of Balkan -- 0.4% for one (in the form of a single segment), and 0.3% for the other (in the form of two segments in approximately that same location as the other sister's one segment).

The excuse that each of my sisters has two X chromosomes seems pretty lame to me, since they're being phased against their father. All 23andMe has to do is ensure that one of each of my sister's two X chromosomes matches our father's exactly. The only place this could be a problem would be at the very ends, but the inconsistency isn't limited to just the ends.

Can anybody account for this?

C J Wyatt III
08-03-2016, 06:41 PM
I think it is a bit of a stretch to try and get any ancestry composition information from the X-chromosome. I would put my efforts elsewhere.

PM me if you wish to discuss further.

Jack Wyatt

08-04-2016, 06:48 PM
Your sisters, of course, have an identical X chromosome to your father, by definition. I think there's not much you can say here other than 23AndMe is inconsistent, which has been established elsewhere too, but it's good to get definitive examples like this.