No Jewish DNA on Ancestry but...

[ada]

My Ancestry DNA result says that I am 95% Eastern European and possibly 5% Finnish/Northern Russian (low confidence region). That's it, nothing else. The European Jewish category says 0-0%.
GEDmatch results reflect that as well (Ashkenazi score on Jtest is 3.67, nothing on Oracle).
I have noticed, however, that my second best match on GEDmatch, with whom I share 31 cM on 3 different chromosomes, is fully Ashkenazi Jewish. With 31 CM shared (the largest chunk being 16 CM) I expect we share a common ancestor. I checked further and it turns out that roughly 1/3 of my matches there are Ashkenazi Jews! They all match me on a few specific locations on 3 different chromosomes (2,20,15, also a small area on 5). Now when I use chromosome painting on gedmatch to look at these particular locations all except one are painted mostly green (east med, west asia) or light green (European jewish). At least these are the colors I share with my many Jewish matches in those areas.
I hardly have any green areas on my other chromosomes.
I attach a picture of one such area - chromosome 20 bet 45 and 54. Top bar is me, below is my best match there who is fully Jewish (>28 on Jtest score) and with whom I share 18 CMs in that area.
Does all of this point to some distant Jewish ancestry or what? What do you think?
I am trying to either debunk or confirm a family legend about my mother's Jewish ancestor.

[josh w.]

My Ancestry DNA result says that I am 95% Eastern European and possibly 5% Finnish/Northern Russian (low confidence region). That's it, nothing else. The European Jewish category says 0-0%.
GEDmatch results reflect that as well (Ashkenazi score on Jtest is 3.67, nothing on Oracle).
I have noticed, however, that my second best match on GEDmatch, with whom I share 31 cM on 3 different chromosomes, is fully Ashkenazi Jewish. With 31 CM shared (the largest chunk being 16 CM) I expect we share a common ancestor. I checked further and it turns out that roughly 1/3 of my matches there are Ashkenazi Jews! They all match me on a few specific locations on 3 different chromosomes (2,20,15, also a small area on 5). Now when I use chromosome painting on gedmatch to look at these particular locations all except one are painted mostly green (east med, west asia) or light green (European jewish). At least these are the colors I share with my many Jewish matches in those areas.
I hardly have any green areas on my other chromosomes.
I attach a picture of one such area - chromosome 20 bet 45 and 54. Top bar is me, below is my best match there who is fully Jewish (>28 on Jtest score) and with whom I share 18 CMs in that area.
Does all of this point to some distant Jewish ancestry or what? What do you think?
I am trying to either debunk or confirm a family legend about my mother's Jewish ancestor.

Yes, it is not certain, but the results could reflect very distant Jewish ancestry, e.g. around a ggg-grandparent. Do you have MO results

[vinnie]

I think the real question is whether someone could have inherited no or very few ancestry informative markers (AIMS), causing the admixture analysis to show no Jewish, versus inheriting segments of shared chromosomes that can be traced to Jewish ancestry. Using myOrigins v1 analyses in tandem with shared cousin matching, I've found that larger percentages of Jewish admixture do not necessarily correlate with a greater number of Jewish matches, at least at lower levels of admixture. For example, in the Jewish and Italian myOrigins Project, one non-Jewish project member with 9% Jewish has fewer in-common matches with another non-Jewish member who has only 6%. Granted, some of the those matches could be coming from a non-Jewish shared ancestor, but so far the data don't support that.

[ada]

Thanks for the input.
For some reason FTDNA does not let me upload raw data so I don't have MO at the moment.
I wondered the same thing. I did seem to inherit considerable chunks of DNA from the ancestor(s) I share with my Jewish matches. How come then I don't seem to have any AJ markers? That would imply that the common ancestor was not Jewish, which is what I assumed. But why do most of these matches match specifically on these rare non-Eastern European areas? And on the few chromosomes which show most AJ admixture (chromosome 15 is around 13% AJ and chromosome 2 around 9 %).
Also, while Ancestry DNA timbered most of my AJ matches, there are some left and we do have 10-18 CMs of DNA in common without having any shared trace regions.
Puzzling.

[khazaria]

This situation reminds me of my mother's chromosome 15 between 88 million and 94 million where she matches 4 members of a particular Polish Catholic family (including a mother and 2 of her children) as well as to another Pole. That is inside Family Finder, not AncestryDNA.

The oldest member of the first family shares 17.6 centimorgans and 2,111 SNPs with my mother on that segment, and Family Tree DNA's MyOrigins 1.0 estimated that person is 100% European, while MyOrigins 2.0 estimates she's 99% European but 0% Jewish. In Eurogenes' Jtest this match's "ASHKENAZI" score is only 0.57% (well into noise territory), the oracle says she's genetically closest to Belarusians, and she isn't given any "AJ" suggestions anywhere in the oracle, not even in mixed mode. She has zero percent scores for the East Mediterranean, Near Eastern, Arabian, and Armenian elements in Eurogenes K36. This match's mtDNA haplogroup has a European origin but some Ashkenazim also have it. These Polish matches have small numbers of Ashkenazic matches altogether. The first family's family tree lists only Christian ancestors and surnames, and their ancestors' records were all in the Catholic Church.

Yet the segment paints in Jtest as West Asian and East Mediterranean! No amount of South Baltic, East European, or North-Central European at all. The segment does triangulate and there are some other valid Ashkenazic matches to it besides my mother.

My mother's MyOrigins 2.0 estimates are:
93% Ashkenazi Jewish,
between 1-2% Sephardi Jewish (1.08% by my revised manual calculations by adding up her Sephardic segments),
and 5% East Europe.

[vinnie]

I think the real question is whether someone could have inherited no or very few ancestry informative markers (AIMS), causing the admixture analysis to show no Jewish, versus inheriting segments of shared chromosomes that can be traced to Jewish ancestry. Using myOrigins v1 analyses in tandem with shared cousin matching, I've found that larger percentages of Jewish admixture do not necessarily correlate with a greater number of Jewish matches, at least at lower levels of admixture. For example, in the Jewish and Italian myOrigins Project, one non-Jewish project member with 9% Jewish has fewer in-common matches with another non-Jewish member who has only 6%. Granted, some of the those matches could be coming from a non-Jewish shared ancestor, but so far the data don't support that.

Clarification: The Italian member with 9% Jewish admixture has fewer matches in common with the Jewish members than the Italian member who has only 6% Jewish admixture.

[ada]

This situation reminds me of my mother's chromosome 15 between 88 million and 94 million where she matches 4 members of a particular Polish Catholic family (including a mother and 2 of her children) as well as to another Pole. That is inside Family Finder, not AncestryDNA.

The oldest member of the first family shares 17.6 centimorgans and 2,111 SNPs with my mother on that segment, and Family Tree DNA's MyOrigins 1.0 estimated that person is 100% European, while MyOrigins 2.0 estimates she's 99% European but 0% Jewish. In Eurogenes' Jtest this match's "ASHKENAZI" score is only 0.57% (well into noise territory), the oracle says she's genetically closest to Belarusians, and she isn't given any "AJ" suggestions anywhere in the oracle, not even in mixed mode. She has zero percent scores for the East Mediterranean, Near Eastern, Arabian, and Armenian elements in Eurogenes K36.

Weird, isn't it? I am also Polish, there is nothing suggesting any Jewish ancestry on GEDmatch but I have 5 segments with only/mostly AJ matches. In 4 out of 5 cases the color we share on these segments is a variant of green (East Med/Ashkenazi/West Asian). The segment on chromosome 2 however looks European - mostly red with minute west asian component.
You seem to be an expert in these things, what's your theory?
I am new to this but having played with Gedmatch for a while it seems to me that chromosome painting on Gedmatch may be unreliable. One of the segments I share with well over 20 AJ matches looks almost completely East Med. However, in the "admixture by chromosome" tool this chromosome shows up as 0% East Med.

I read about the new Ancestry DNA test picking up minute traces of Native ancestry dating back to 17 or 18 century. The AJ ancestry should be even easier to detect. I even get 0-1% south asian reading but 0-0 Ashkenazi. This leads me to believe that the connection, while real (there are just too many solid matches on different chromosomes/segments) must be through a non Jewish ancestor. But I don't know.

[khazaria]

You seem to be an expert in these things, what's your theory? ... This leads me to believe that the connection, while real (there are just too many solid matches on different chromosomes/segments) must be through a non Jewish ancestor. But I don't know.

It's possible. The EAST_MED and WEST_ASIAN elements in Eurogenes Jtest and EUtest and Eurogenes K13 are different from the similar elements in Eurogenes K36. It is not rare for full West Slavs and East Slavs to show EAST_MED and WEST_ASIAN in Jtest and EUtest and K13 due to ancient ancestral migrations from south to north, but at the same time for those same people to score 0 percent of any region from the Middle East in Eurogenes K36. So I cannot automatically assume that every East Mediterranean segment is of Jewish or Levantine origin.

I am new to this but having played with Gedmatch for a while it seems to me that chromosome painting on Gedmatch may be unreliable. One of the segments I share with well over 20 AJ matches looks almost completely East Med. However, in the "admixture by chromosome" tool this chromosome shows up as 0% East Med.

I have had good success with chromosome paintings, especially for Sephardic Jewish segments, which more often than not show up as pure East Mediterranean or Middle Eastern segments. I think that is reliable. Eurogenes' creator does not recommend using the numerical admixtures by chromosome. He does recommend the visual paintings by chromosome.

What I think is a small part of my mother's ancestry is ambiguous and the genetic evidence did not necessarily support the existence of a recent Polish ancestor but still could, whereas in my father's case I was able to find unmistakable evidence for Polish ancestry on Slavic segments with pure Slavic matches and I believe the genetic evidence shows he had a 6th great-grandparent who was a Pole.

[ada]

I have had good success with chromosome paintings, especially for Sephardic Jewish segments, which more often than not show up as pure East Mediterranean or Middle Eastern segments. I think that is reliable. Eurogenes' creator does not recommend using the numerical admixtures by chromosome. He does recommend the visual paintings by chromosome.

The most I share with an AJ match is 31 CMs total. The longest segment I share with an AJ match is 18.7 CMs. Among my top 50 matches on Gedmatch I have counted 24 AJs. /I am an immigrant, with only distant family here, which explains a lot. Still, this over representation piqued my interest and I immediately remembered my mother's story about her supposed Jewish ancestor/.
Based on the above, I think that a common ancestor existed, and lived within the last 10 generations. Do you think that's a reasonable assumption?
If I wanted to determine if that ancestor was Slavic European or European Jewish which calculator do you think I should use to paint my and my matches' chromosomes? Jtest? Eurotest? K13? Other?
Thanks everyone!

[khazaria]

• When I do the segment area paintings, whenever a match I'm studying is not only in Family Finder but has also joined GEDmatch, I specifically use the Admixture screen's feature "Paint differences between 2 kits, 1 chromosome" and compare multiple pairings of holders of a segment and narrow down to the element(s) that always show up in common across all pairings, and I start with Eurogenes' Jtest.
• When the "ASHKENAZI" element shows up, its ethnogeographic breakdown can be found by repeating the paintings with Eurogenes' EUtest.
• Other calculators can help provide a "second opinion" about the ethnic origins of a segment if it looks confusing or inconsistent.

Yes, I do agree that your shared ancestors with those Jews likely lived within the past 10 generations because of some of those substantial segment lengths you mentioned.

When I see only Middle Eastern elements on a segment shared with Jews, my inclination is to believe that the shared ancestor was Jewish, even when it seems against all odds judging by the overall ethnicity estimate of the non-Jewish match(es). There are some Hispanic people in Family Finder who hold a definite Jewish segment or two even though they score 0% in MyOrigins' Jewish categories.

On the other hand, we know that there were some Poles who had Jewish descendants because there are some (small numbers of) Ashkenazim with the West Slavic Y-DNA R-M458 or with mtDNA haplogroups that are likely of East European origin.

[JerryS.]

follow the GEDmatch results, they use the real science of the chromosome, not some "estimate" based on a limited population pool sample.

[ada]

follow the GEDmatch results, they use the real science of the chromosome, not some "estimate" based on a limited population pool sample.

Yes, I think you are right. These tests are unable to detect and/or correctly assign very minor admixtures.
I have since managed to transfer my ancestry raw data to ftdna. I only got 48 matches at ftdna - people with obviously Polish roots and Ashkenazi Jews, in roughly equal numbers. Per "my origins" I am virtually 100% European (well 98% but the rest is unassigned) and my best AJ match is listed as 0% European. Yet we share a 20 CM segment and a further 6 cm on X, it's obvious than this test is missing something.
It would appear from chromosome painting the AJ matches and me share East/West Med or related colors so that ancestor was probably, maybe Jewish. I only have Jewish matches on those segments. Since no test is picking this admixture I think it must be very ancient.