PDA

View Full Version : Where is the cutoff between general Ashkenazi Jewish and non-Jewish admixture?



Riverman
07-09-2020, 01:06 PM
I asked the question in another thread, but got no answer and found nothing conclusive about it:
Looking at what companies like 23andme or Myheritage assign, most people with 4 Ashkenazi grandparents score especially on 23andme something like 99 % AJ. How far back would a recent admixture be recognised by the algorithm? If someone had 3 non-AJ ancestors in the 18th century, would that stick out? They wouldn't have been part of the bottle neck and being recent admixture instead, but still back far enough for being rather widespread and integreated in the local AJ gene pool. Yet they wouldn't be "common Ashkenazi", but restricted to only some locality and families. Where would those with experience for samples from different AJ groups draw the line between what's being commonly subsumed under AJ and what sticks out as newer admixture? Does it depend on the samples a calculator uses only? Like if the recent admixture being part of the reference, it wouldn't be recognised as "new admixture" in one case, while if the reference lacks it, it would? That way by having very large AJ samples, any newer admixture even from the 19th century might disappear?
There surely was admixture from the 16th to 19th century, as small or big as it might have been and the later it happened, the less widespread and generally Ashkenazi it can be.

StillWater
07-09-2020, 05:14 PM
You're effectively asking how 23andMe works and how inheritance works. All are relevant points to consider. Their algorithm + large reference group is probably why Ashkenazim with many Sephardic segments still get the same AJ scores as any other AJ. I also think that based on what I've seen partial Krymchaks receive on 23andMe, that their "method" (algorithm + reference set) demonstrates that Crimean Jewish admixture does exist in Ashkenazim.

eolien
07-13-2020, 01:52 PM
According to their documentation, they (23andme) use blocks of 0.2 cM. This is pretty detailed. If they were smart they would use German Jews but i am not so sure. They might have used multiple AJ populations for reference which makes their method somehow more conservative (AND vs OR approach)

StillWater
07-13-2020, 02:21 PM
According to their documentation, they (23andme) use blocks of 0.2 cM. This is pretty detailed. If they were smart they would use German Jews but i am not so sure. They might have used multiple AJ populations for reference which makes their method somehow more conservative (AND vs OR approach)

Remaining German Jews are unlikely to have all the old Rhineland Jewish segments.

Riverman
07-13-2020, 02:57 PM
Remaining German Jews are unlikely to have all the old Rhineland Jewish segments.

And possible more recent admixture as well.

Seabass
07-13-2020, 03:10 PM
How much Ashkenazi would a full Krymchak get? I seen 3 Crimean Karaites, they each get between 0-1% Ashkenazi Jewish. Frankly very TINY compared to other Western Jews. The three full German Jewish relatives (allegedly) I share with get 99.9, 99.9 and 99.3 with 0.7% WANA. I wish I knew how much Ashkenazi some Italian Jews could get.