Wednesday, November 26, 2014

Genes From My Father

My father's name
This Wednesday evening, the fifth of Kislev, my father will have been gone thirty-four years. His name is Eliezer Yitzhak. Aunt Helen called him "Leizer Yitzchok."

Based on his initials, Lawrence Irwin Pickholtz, my mother often called him "Lippy." (Actually, "Lippy dear.") What my mother didn't realize is that the standard Yiddish kinui (=nickname) for Eliezer is Lipa.

My father was named after his great-uncle Lajos/Ludwig Bauer who was a Hungarian government minister at some level and who was probably living when my father was born. He would have been two days short of forty-eight.

Simon Bauer's 1862 marriage record, naming his father Lasar
We know that my third-great-grandfather was Lasar Bauer but we have no specific knowledge that either he or Lajos/Ludwig had the second name Yitzhak. I missed the chance to ask my grandmother about that.

It is quite possible that the Yitzhak comes from my grandfather's side, as my second-great-grandfather was Isak Fischel Pikholz. My grandfather had an older brother named Eisik Fischel who died at nineteen months. So my father's second name may have come from there. Otherwise none of the descendants of my great-grandparents carry his name.
Eisik Fischel's birth record (adjusted on the page)
My father's DNA
A few weeks ago, GEDmatch came out with some new analytical tools that require a subscription. One of them - Lazerus - uses descendants and relatives who are not descendants to reconstruct the DNA of deceased people.

You enter descendants in Group 1 and non-descendants (siblings, cousins etc) in Group 2. Lazerus then assigns the reconstruction any segments that match people in both groups. The assumption is that if my aunt and I have a match, my father must have had that same segment. (I am probably over-simplifying here.)

As with all DNA issues, the more actual test results, the better.

The reconstructed kit, if large enough, can then be compared to other kits, the same way any other kit can be.

I have been talking about doing this for my great-grandparents since last summer, using the tools developed by Kitty Cooper. But Lazerus is easier, almost automatic. This is good, but also problematic, for reasons I'll explain below.

Running Lazerus for Leizer
But before I tried Lazerus on my great-grandparents, I decided to run it on my father. (The irony of using Lazerus to reconstruct the DNA of someone named Leizer was not lost on me.)

I hope the explanation is not too technical and that I make myself understood, to both the laymen and the citizen-scientists among my readers.

I have three descendants for Group 1 - two of my sisters and me. (A third sister said she will test, but hasn't done so yet.) For Group 2, I used the kits of my father's sister and brother, Aunt Betty and Uncle Bob, and their first cousin Herb. (There is one other living first cousin and I am working on him to test.)

There are several others a bit more distant whom I could have used, but I decided to start conservatively in order to reduce the possibility of endogamy. If, for instance, I were to include my second cousin Terry or my father's second cousin Bruce in Group 2, I would be taking the chance that their mothers have some kind of match with us via my mother and that would find its way into my father's Lazerus kit where it doesn't belong.

Herb and Kay (Surely no one in our family has ever referred to those two cousins together!)
The way I set it up, the only person who could have that endogamous match would be Cousin Herb. So to check that out, I looked for segments where Herb matches my first cousin Kay, whose mother is my mother's sister. There are five such segments.
Matching segments - Herb and Kay

On the first segment on chromosome 3, Aunt Betty and Uncle Bob both match my sisters and me, so we don't need Herb's match.

On the second match on chromosome 3, Aunt Betty and Uncle Bob do not match us, so this segment is disqualified from the Lazerus because it is based solely on Herb's match with my mother's side.

On chromosome 7, we have the same thing. That segment has to be disqualified.

The segment on chromosome 20 is covered by Aunt Betty and Uncle Bob.

The segment on chromosome 12 is more complicated. Part of the Herb-Kay match is covered by Aunt Betty and Uncle Bob, but a bit at the beginning is not.

I cannot be precise, but it looks like I should remove 2100-2200 SNPs from the Lazerus kit that was generated for my father. If there were a way to do so. (GEDmatch does not offer a mechanism to remove - or add - specific segments.)

The total Lazerus kit has 666,901 SNPs, so we are talking here about less than a third of one percent, so the Lazerus kit for my father is essentially free from endogamy.

I'll spare you the math, but the kit has about 62% of a full set of twenty-two (X2) chromosomes. Others may have more or less.

A second trial kit
I then ran a second kit for my father, but I added six other people to Group 2 - a child of each of three of his first cousins and three more distant cousins. Those have a higher likelihood of introducing segments that should not be there. That kit has about 1.24% more SNPs than the first kit. That increase hardly seems worth the bother, so I decided to stick with the first kit.

Very small segments
Let me clarify one other point. Generally, when we compare peoples' DNA to one another, we ignore the really small segments. Where we draw the line is a matter of personal preference, but most choose somewhere between 6 and 10 cM as a lower limit and reject anything smaller. In this case, I decided to accept all the small segments since we are talking about known close relationships, where the small segments would have certainly been passed on from the previous generation. It doesn't matter if they are not relevant to three-four generations back.

In any case, it would not be significant. I have only one segment smaller than 7 cM  matching the Lazaerus kit. My sister Sarajoy also has only one and my sister Amy has three. Not much to be concerned about there.

Analyzing the Lazerus kit
So how do I analyze my father's kit once I have it? And why is it significantly different from what I can do with the kits of Aunt Betty or Uncle Bob? I am not sure about the second question, but I can show you some of what I did.

I ran my father's kit - remember it's not a full kit - in GEDmatch and of the fifteen hundred matches, I pulled out the twenty-five that are part of my own project.

Those twenty-five break down as follows:
The descendants of my great-grandparents - 9
Kwoczka cousins (my g-gm's side) - 2 (Bruce and Pinchas)
Descendants of my g-gf's sisters - 1 (Ralph)
Other Skalat Pikholz - 8
Other Pikholz - 3
My mother's Jaffe cousins - 2
I ran a chromosome browser with those twenty-five. What I would really like to do is to see how the other Skalaters might fit together, but I am not really sure how best to do that.

In any case, here are a few of my father's matches, by chromosome. The colors indicate the quality of the match for each segment.

The order of the matches as displayed is different for each chromosome. (Note to GEDmatch: It would be easier to read if the order stayed the same throughout.)

On chromosome 1, as on most of these charts, the top five are my sisters and I, Aunt Betty and Uncle Bob. The order of the basic five is not important. We do see that they line up pretty well with one another.

Line 6 is my second cousin Rhoda and 7 is my father's cousin Herb. Line 8 is Micha from Rozdol. Line 9 is Charles, a Skalat cousin whom I wrote about a few days ago, who should be moderately close to us. Line 10 is the putative great-grandson of Uncle Selig.

There is a nice set of overlapping matches on lines 13-17. Line 13 is Ralph, which should identify this segment as Pikholz rather than Kwoczka. Jane is 14, my cousin Lee is 15, 16 is Gene from the Irene family and 17 is Lloyd. (Some of these names should be familiar to you from my previous post, where I identified them in detail.)

Lots of other small scattered matches.

In chromosome 3, the first eight are the basic five plus Herb and my cousins Terry and Rhoda, followed by my cousin Lee on line 11. In between, on lines 9 and 10 are the two Jaffe cousins on my mother's side. I do not believe this is an error of endogamy, but rather a legitimate very old connection between the families. One or both of the Jaffes appear in other chromosomes as well.

Line 12 is from Uncle Selig.

There seems to be a bit of grouping at the bottom right, involving a group of Skalater Pikholz  descendants. Here too we see Micha on line 19. He seems to have some round-the-back connection to the Skalat Pikholz families.

On chromosome 7, we have something interesting. Directly after the basic five comes my second cousin Lee and there is a significant segment on the far left that includes mainly those six people. Lee's grandfather is my grandfather's brother, but our grandmothers are sisters. So my guess here is that this segment - which has no other Pikholz matches of note - is from our grandmothers' side - the Rosenzweigs (or Bauers or Sterns or Zelinkas).

On chromosome 18, lines 7 and 10 are the Kwoczka cousins Bruce and Pinchas. We can safely say that those segments come from my great-grandmother's Kwoczka side rather from my great-grandfather's Pikholz side. We have the same phenomenon on other chromosomes, particularly for Bruce, who is my father's second cousin. (Pinchas is one generation later.)

 Finally, chromosome 19, where we have the basic five followed by Herb and Lee, taking up almost the entire chromosome. The important one here is line 9, where Ralph claims that segment as Pikholz, rather than Kwoczka.

The longish segment on line 8 is Gene from the Irene family. His third cousin Bonnie is on line 11.

All this is mostly experimental. The real test will come from a look at the great-grandparents. Lazerus or no, it still may make sense to copy some of these segments into Kitty Cooper';s Chromosome Mapper for the great-grandparents.

Moving on to the great-grandparents
As I said, what I really want to do is Lazerize my great-grandparents, Hersch Pickholz and Jutte Lea Kwoczka. Both parents of Hersch Pickholz are named Pikholz and if I can recreate a significant amount of his DNA, I can compare that to other Skalat Pikholz families, as well as the Rozdol Pikholz descendants and the set of non-Pikholz who have joined our project.

I discussed this subject three months ago when I wanted to do this using Kitty Cooper's Segment Mapper. This was a difficult task and in any case, there was no easy way to compare other people to the finished product.

This is no simple matter because although we have tests for nine descendants of Hersch and Jutte Lea, we have only one relative on his side and two on hers to help distinguish between them. And there is the matter of endogamy - having to try to figure out whose matches might be from another direction.

I am also thinking about whether it is right to include my father's kit in my ggps Lazerus - or whether it makes a difference in the results.

Lazerus kits are easy to compare, so that problem is solved, but the problem of distinguishing between my two great-grandparents remains. And checking for endogamy is both an uncertain task and a tedious one.

Also note that Lazerus looks for matches between the segments in Group1 (the descendants) and those in Group 2 without our having any control. When I was doing it with Kitty Cooper's Segment Mapper, I could look at each potential match and decide if it was good enough for me.

The great-grandparents is the next big project on my list.

My thanks to Curtis and Jon at GEDmatch, to Jim Bartlett and to Rebekah Canada for comments along the way. And, as always, to all the cousins who have tested.

Housekeeping notes
FamilyTree DNA is having a sale on DNA tests and upgrades.   A good explantion of all the options is here. Order here.

One Pikholz has already ordered a Family Finder at the sale price..

There is now a dedicated Facebook group for GEDmatch Lazarus tool.

No comments:

Post a Comment