Wednesday, October 22, 2014

Using GEDmatch

Anyone who is serious about doing DNA for genealogy uses Sooner or later. It's been later for me but I am finally here.

GEDmatch is a free site run by two volunteers, which provides tools for analysis of autosomal DNA match results. People who have tested with one of the three testing companies upload their results to GEDmatch - it can be done anonymously - and the GEDmatch tools allow comparisons among results from all three testing companies, in addition to advanced analyses that even FTDNA does not offer internally.

GEDmatch is meant to be intuitive, but my own intuition on these things is not what it once was, so I have been hesitant about getting involved. (I was disappointed that my GRIP course did not include any hands-on GEDmatch exercises.)

Nonetheless, now that I have been interacting with some of the big names in the field, I felt that I was ready to have a go at it.

Matches with Miriam
My first experiment was for a woman named Miriam who tested with Ancestry and who showed many matches with my families based on a look-up on GEDmatch. Here are results from her eighteen interesting matches with us. There are five more among the Rozdol Pikholz but there was nothing significant there and I left them out for space reasons. You will probably need to click on the image to see the actual numbers. (The splotches of color I added myself and will explain below.)
The five tables on the left are Miriam's matches with me, my two sisters and my father's sister and brother.

The first two in the second column are second cousins on my mother's side (first cousins to one another). The bottom two are descendants of Peretz Pikholz (~1820-1873). 

The top two in the third column are my father's first cousin Herb and my second cousin Rhoda, followed by my two cousins on my great-grandmother's Kwoczka side.

The top three in the column on the right are descendants of Nachman Pikholz (~1795-1865) and the two at the bottom are indeterminate Pikholz/Bernstein cousins.

I then added a splotch of color to the matching (or near-matching) segments so Miriam could see her matches more specifically. For instance, my match with Miriam on chromosome 4 (starts at 97,112,821 and ends at 111,328,469) is nearly identical with her matches with one of my sisters and my aunt. I marked them in green.

My two sisters and my aunt have a match on chromosome 9 and I marked them in red.

My aunt and my uncle share three segments, one of which they share with one of the Kwoczka cousins and another with both him and Rhoda.

What does all that tell us? First it says that the grey and purple segments on chromosome 9 come from the Kwoczka side and that Miriam and my ggm have a common ancestor.

It is hard to say more about the matches with my sisters, my aunt and uncle and me. We have no one to test on my grandmother's side and there are no matches shared by the others on my grandfather's side.

Miriam has two matches with the cousins from my mother's side, but not with us. That may mean that Miriam has something in common with their Jaffe grandfather. I told Miriam Sunday morning, when we Skyped, that it would be nice if GEDmatch had a way to see who else in their database matches those two specific segments. Little did I know!

Finally, Miriam has a match with two of the descendants of Nachman. Those two are second cousins to one another and that match is likely from their "other side."

Aunt Betty's X chromosome
I have not been doing anything regarding X chromosome matches. This is one of the sex chromosomes - women have two of them and men have one plus a Y. Men get their X from the mother and women get from both parents.

Later Sunday, a woman I have been speaking with on Facebook - Kathy, who I think is in New Zealand - posted something about X chromosome analysis using GEDmatch and as a result of our ensuing discussion, I decided to have a look at Aunt Betty's X matches within the Pikholz families.

Here are her ten matches.

As in the case of Miriam's matches, I marked the identical or near-identical segments with color.  The blue segment shared by Aunt Betty, Lee and Pinchas (a Kwoczka cousin) almost has to be on the Kwoczka side. It cannot be on the Pikholz side, as my grandfather did not receive any X DNA from his father, only from his mother.

In fact, Aunt Betty cannot have any Pikholz X DNA, so her matches with Herb, Terry and Rhoda are either Kwoczka DNA or something from my grandmother's side via some unknown (and unimagined) common ancestor. That is not as trivial as it sounds, because the red segment is shared by Rhoda, Aunt Betty and my two sisters. Any X DNA my sisters received from our father must have come from his mother, as he received no X DNA from his father. (The same is the case for the green segments that my sisters share with Uncle Bob.)

But it does clearly mean that Rhoda has a connection on my grandmother's side and it must be on Rhoda's mother's side.. (Rhoda's father could not have gotten from his father, so the connection is not that way.) I know nothing about Rhoda's mother's family, but Monday evening she and I had a furious exchange of emails which included my finding quite a few documents for her mother's family.  My grandmother's known ancestral names include Rosenzweig and Zelinka from Slovakia and Bauer, Stern and maybe Grunwald from Hungary. I was hoping to find one of those surnames in Rhoda's mother's family.

I ordered the New York marriage record for Rhoda's grandparents and in the meantime she told me that her maternal grandfather was from Galicia. The documents I found this week show that Rhoda's maternal grandmother's parents were from the Humene area in Zemplen County, Slovakia. The surnames there are Eichler and Schonfeld. Perhaps something more useful will turn up later.

In the meantime, Sunday afternoon, I saw a Facebook post by Blaine Bettinger, one of the lecturers at my GRIPitt course, saying that GEDmatch had announced four new tools in a category they call "Tier1." There was no meaningful explanation of what those four would do, but the idea is that you buy a subscription at $10 a month for access. The basic GEDmatch will remain free.

So far, Blaine has blogged about a tool they call Lazerus, which recreates the DNA of ancestors based on the tests of living descendants. As my readers know, I have been trying to do this for the last six or eight weeks, using Kitty Cooper's mapping tools which Kitty herself ran with soon after. I am thinking this Lazerus tool (Blaine has already turned "Lazerus" into a verb) may do the trick, though I am concerned that it may overlook the   endogamy problem. That could result in recreating the ancestors incorrectly, so the process will have to be monitored.

When they receive my subscription, I'll be able to have a look. Meantime, here is how Kitty describes Tier1:

I will also be looking forward to seeing what Jim Bartlett has to say about this. He is not an endogamy expert, but he is very very good at using these tools and in particular knows how to explain it to the layman. He gave an excellent talk at the recent International Genetic Genealogy Conference which I purchased.

I also very much  enjoyed the talk by Tim Janzen MD from the same conference and our subsequent discussions about chromosome mapping. Tim is considered one of the top people in the field and he has convinced me that we should have a bit of testing with the other two companies, 23 & Me and Ancestry. Both those companies have issues with orders from outside the US, so I asked Aunt Betty do those tests. In any case, her DNA is one generation better than mine  anyway. Tim also said I should get some tests from one of her sons and we are working on that.

Kitty herself has blogged about the triangulation feature of Tier1.

If you recall, I mentioned above that I told Miriam that she needed to search for people who match specific segments of my Jaffe cousins and that is another feature that we find on Tier1.

Now that the holidays are behind us, I hope to get back to my regular Sunday morning schedule, either next week or the week after.


  1. Nice post Israel! You can easily spend a lot of time at GEDmatch. You probably won't be surprised to hear that Jim Bartlett had a hand in developing Lazarus. I haven't heard him chime in publicly since Lazarus launched, but I'm looking forward to any information he can share about the new tool.

    1. Thank you, Blaine. I may be old-guy-slow, but I am going to break tgrough some of this endogamy stuff sooner or later.

      Knowing some of the big people in the field helps immensely.

  2. This comment has been removed by the author.