Some of the least reliable information provided by DNA tests are the ethnicity percentages, also called admixtures. After getting my and my uncle’s DNA test results, I posted my ethnicity results. Our Ashkenazi Jewish percentages came in at 79% and 89% at FamilyTreeDNA and my result came in at 83.8% at MyHeritage DNA (which you’d think must be accurate because of the extra decimal). When FamilyTreeDNA “improved” their Ethnic Makeup algorithm a few months ago, my Askenazi Jewish percentage went up from 79% to 92% and my uncle’s went up from 89% to 96%. Better, but still not the 100% I expected.
At the best, these percentages are probably within 20% of their true value. If you’re up in the 80% range, you know you’ve got a lot of that ethnicity in you. But what if you’re in the 20% range or 10% range. Is that significant, or is it just some ethnicity mapping problem and you’re actually zero? People get surprised to find 10%, 5% or even 1% Ashkenazi Jewish in their DNA when they’re not expecting it. They want to know if that’s real.
What we need is another test to verify whether you’ve got some DNA connection to this endogamous group.
There is the JTest admixture calculator you can use at GEDmatch that is hidden deep in their Admixture utilities under the Eurogenes project. But it’s caveat emptor and you likely need a PhD degree to properly interpret the results. I’m 27.99 which supposedly is a significant Ashkenazi Jewish number, but this means nothing to me.
What we need is a simpler test. And today I got inspired by a thread in the Facebook Group: DNA Help – Your Jewish Journey. One member of the group mentioned that she matches 26 of Israel Pickholtz’ test kits. Another said they they think there’s a t-shirt you can buy that says “How many Pickholtz DNA kits do you match?”. Israel has 100 or so kits. When Lara Diamond (who has about 50 kits) entered the conversation, people then started commenting on how many of Lara’s kits they were matching.
When I first tested my uncle just over a year ago and got his results back, I looked through his 7,016 matches that he then had, and found that Israel was his 2,003rd match and Lara was his 2,364th match. I had earlier purchased Israel’s book, Endogamy: One Family, One People and found lots of matches to the people he tested. Similarly I found matches to Lara’s kits. So this thread on Facebook was nothing new to me.
Many years ago, I heard somewhere (source unremembered), that there was a simple way to estimate the number of Jewish people in any city. All you had to do was take the city’s phone book (for the people who remember what a phone book is) and go to the pages with the surname: Cohen. Then simply count the number of Cohens in the phone book and multiply by 100 and that is a good estimate for the number of Jewish people in that city.
All right. Let’s do something similarly simple.
The Pickholtz-Diamond Index of DNA Ashkenazi Jewishness (PD Index) is determined as follows:
If you have a FamilyTreeDNA account, go to your Family Finder matches and download them to Excel. Search for “israelp@” and do a Find All. Write down how many you find. Search for “laradiamond@” and do a Find All. Write down how many you find. Average the two numbers and that is your PD Index.
You can also do it at GEDmatch. Do your ‘One to many’ matches (don’t change the default values). Count the number of “israelp@” and “laradiamond@” that you find. GEDmatch only shows your top 2000 matches but don’t worry about it.
While you are doing this, you might want to note If any of your matches have a largest segment of at least 20 cM. If so, you could think about contacting either Israel or Lara about that and see if you can connect. See Lara’s article: A Technique for Endogamous DNA Using GEDmatch.
So here’s a table I’ve put together to illustrate with my matches:
Thus, a PD Index value of 20 on FamilyTreeDNA and 16 on GEDmatch is a good indicator of healthy portions of Ashkenazi DNA.
As a check, if I determine the PD Index for Patricia Gerber (see my last post), she shouldn’t have any Ashkenazi in her. And her PD Index at GEDmatch in fact comes out to a big zero. Neither Israel nor Lara show up in Patricia’s match list. And looking back at that thread on Facebook, one person who has one Jewish grandparent said they found 3 connections with Israel and 2 with Lara.
The sad part of this all: So far, not a single common ancestor between myself and Israel or myself and Lara have been determined from our DNA connections.
I look forward to meeting up with Israel and Lara again at the IAJGS Conference in Orlando in three weeks time.