Sunday, February 23, 2014

COUSIN MARLA

After three consecutive weeks of blogging about Pikholz DNA, I planned to change the subject to something else - anything else.

Then I heard from Marla.

I had tweeted about having been accepted for the GRIPitt course in Practical Genetic Genealogy and Marla congratulated me on that. Having recently looked at the MtDNA matches for Pikholz descendant Amos (of the only Pikholz line that is all female) and seen Marla's name there, I decided to see whom she matched within our group.

Marla matches fifteen out of the twenty-six Pikholz who tested, but oddly enough Amos is not one of them.

The most interesting of those matches was with my father's cousin Herb, my father's sister Aunt Betty and me. Using Herb as the base, Marla matches all three of us on chromosomes 1, 16 and 22 (5.46 cM, 5.47 cM and 7.68 cM) and has another match with just Herb and Aunt Betty on chromosome 11 (5.78 cM). She did not match my second cousin Terry.

I reported this to Marla and she told me that her brother, her mother and her mother's brother also tested. Marla's brother also has fifteen matches - eleven of them are the same as Marla's and each has four that the other does not. So that's nineteen between them. Marla's mother has seventeen matches with us and the uncle has sixteen. All four have now joined our group.

Here is how Marla's group of four lines up with our three. Considering that we are defined as "remote" and cannot be closer than six generations, those seem to be very nice matches.
So I am guessing that Marla's family and mine have their Most Recent Common Ancestor (MRCA) six-seven generations ago. Her family surnames on her mother's side do not appear in our family.Those names would be Uszerowicz, Lewkowicz, Organek (previously Organkiewicz), Herszkowicz, Szlamowicz, Lichtman, Szapsiowicz, Baltman, Baran, Rubinow, Abramahow. (Note to self - clarify with Marla that all these "...wicz" names are documented surnames, not just patronymics.)

Lloyd has the same match on chromosome 11 with Marla's mother, but not with Aunt Betty. Dalia - who is Lloyd's third cousin on that same Pikholz side - does not.

Jacob also has the chromosome 11 match with Marla's mother, but not with Aunt Betty. The other three from Jacob's group do not.

I did not discern anything significant between Marla's group and the other Pikholz matches.

What characterizes all the Pikholz who match Marla's family on chromosome 11 is that they are Skalaters and that they all come from our matching Y-DNA group. Perhaps a Y-DNA test of Marla's uncle would help here, though a negative finding would not prove or disprove anything.

I do think that of all the non-Pikholz families who match a large number of Pikholz descendants, Marla's family is the one which appears a match within six-seven generations to a particular set of Pikholz families. So with that, I'd like to extend a tentative but hearty welcome to Marla and the other members of her mother's family. You really appear to be cousins, nearly in reach.

For next week, you'll be pleased to know that I already have another topic that has nothing to do with either Pikholz or DNA.

Housekeeping notes
The left hand that appears in the right-hand column under the "followers" means that I am now registered on the Geneabloggers website. The fellow who runs it is a well-known genealogist and lecturer.

The first of the four books I need to read in preparation for the GRIPitt course has arrived.


Sunday, February 16, 2014

VISITING THE X-CHROMOSOME

So until now, in the Pikholz DNA Project...
I have looked at the Y (male line) chromosome of some of the Pikholz families and we see that three of the Skalat families have a perfect match to one another at 37 markers. This tells us that these three families have a common ancestor most likely six or seven generations back. (They cannot be closer, as we know the five most recent generations.)

The Y tests also told us that certain other family members do NOT have a male line that matches the rest of us.

We did a MtDNA (female line) test on a member of the one Pikholz family that has only an all-female line. That showed a few good matches with non-Pikholz tests, but nothing good enough to work with.

And we have been looking at autosomal DNA tests, as I have discussed before on several occasions. Those are the twenty-two pairs of non-sex chromosomes. Most of the results have been ambiguous to this point.

The X chromosome
So now I'd like to say a bit about the X chromosome. Men have one, which comes from their mother. Women have two, one from the mother and one from the father. The one a woman gets from her father is the same one he got from his mother, so essentially, women have one from the paternal grandmother and one from the mother.

Genetic genealogist Blaine Bettinger put together these diagrams which show how the X chromosomes of men and woman are made up. (Pink represents female ancestors, blue male.)

The circle in the center is the male. His X is from his mother. She in turn gets from her two parents, who get from his mother and her two parents. And so on. The number of ancestors in each generation increases according to the Fibonacci Sequence.

The percentages in the illustration assume that a woman passes on half from her mother and half from her father. but this has been found to be incorrect, as Roberta Estes demonstrates. The mixture that she passes on from her parents can be 50/50, but it can also be 90/10 or 73/27 or anything else.


On the right we see the female chart from the circle in the center. Her right side is the same as that of a man's, all of it coming from the mother in exactly the same way.

But she also has an X from her father and that is what appears on the left.

Family Tree DNA, the test company our project uses, began to include the X chromosome in its results about a month ago.

(Let me emphasize that the X chromosome is not the same as MtDNA, though both involve the mother.)

The surname that will not go away
The first person I looked at was Mark. He is a non-Pikholz from a Skalat family, who matches many of us, and I wanted to examine his theory that he matches us on his mother's side.

He matches six of us on the X chromosome, in two distinct groups, both of them nearly perfectly overlapping. A seventh match (Micha) overlaps no one else, so I am ignoring that.

In the first group, Mark matches Dalia, Gadi and Robert. Mark's matches with Gadi and Robert appear identical. As it happens, both Gadi and Robert are Rozdolers, but these are X chromosomes, so they would be from their mothers' (non-Pikholz) sides.

The match with Dalia is nearly identical. In her case, the match could come from either her Pikholz father or her non-Pikholz mother. Dalia is a Skalater.

The second group has much smaller colored bits but is much more interesting genealogically. Herb and Betty are first cousins. (Betty is my father's sister.) Their matches with Mark are identical. (I could not appear there because my X is from my mother's side. My second cousin Terry could, but doesn't.)

The third person in this second group, Miriam, is a Rozdoler. Her match cannot be from the Pikholz side, as that would be her father's father's side, and her father received his X from his non-Pikholz mother.

So I looked to see what I know about Miriam's mother's side and it turns out that Miriam's mother's maternal grandmother is Miriam WACHS. Well, we certainly know the name Wachs. That was the family I discussed in my very first blog post.

We know that Herb has Wachs in his ancestral backgound, on his father's side. There is a persistent tradition, confirmed by Wachs descendants from Pittsburgh, that there is some sort of connection between them and my own Pikholz family from Pittsburgh. Unfortunately, no one has a clue what it is.

Mark tells me that his paternal grandfather had a cousin who married a Wachs in Tarnopol, but that shouldn't matter because that Wachs would have no common ancestry with Mark and besides, the connection is on Mark's father's side which doesn't affect the X chromosome.

So I haven't a clue how to resolve this.

Robert
Next, I decided to look at X matches for Pikholz males - generally their non-Pikholz sides. I wanted to see if some of the non-Pikholz who matched many of us did so on our mothers' (non-Pikholz) sides. When I did this for Robert, I find something remarkable. He matched three Pikholz on his mother's side, two of them men, which meant on their mothers' sides as well.


Not only that, but as you can see on the right, those matches overlap with one another.

Note that I had inadvertently set the level of testing to 1 cM (centiMorgan), which would include very small segments, but as you can see, these are not very small segments, so that appeared to be irrelevant. At least that's what I thought.


Since one of the excuses the experts use to explain this kind of thing is that the matching segment is too small to be meaningful, I had a look at the precise sizes. And here there was a surprise. The longish green match between Robert and Dalia turns out to be two nearly adjacent matches, the first 1.52 cM, the second 5.23 cM. Robert's matches with Micha and Gadi are 5.83 cM and 2.16 cM respectively.  The experts say to ignore bits less than 5 cM, so we can ignore Gadi here entirely.

I'm not sure what the close matches among Robert, Micha and Dalia can tell us, but we can certainly see that the size of the colored bits on the chromosome browser is very misleading. Gadi's match with Robert looks to be bigger that Micha's, but in fact it is less than half the size.

Another odd thing here is that Robert Gadi, Dalia and Micha match in the same place as Mark, Gadi and Robert but Micha and matches Mark in a different place on the X chromosome.

So I went back to Mark
The actual numbers for group with the larger colored matches with Mark are Dalia -1.29 cM. Gadi and Robert - 1.41 at exactly the same place. These numbers are supposedly too small to be on interest, but it is hard to ignore the precise overlap of the three men and the near-precise match with Dalia.

The actual numbers of the very small colored segments in Mark's second group are Herb and Aunt Betty 4.65 cM in exactly the same place and Miriam 8.2 cM. Miriam's match with Mark includes the entire segment of A. Betty and Herb. These are much more significant numbers that the ones that appear larger.

Now I have to figure out what to do with them.

But I have a plan 
I really have to learn more about how all of this works and to that end have successfully enrolled* in the week-long Genealogical Research Institute of Pittsburgh course in Practical Genetic Genealogy to be held in July in Pittsburgh.

I am hoping that that experience will make me better equipped to analyze all these test results.

I also hope that the lecturers do not end up saying that Ashkenazi Jewish (what they call AJ) DNA is too difficult to do well, because the population is too deeply inbred.

* Successful enrollment was no small matter. When registration opened last Wednesday at 7 PM my time, they had so many applicants that the server crashed. They were finally up at 7:09 and when I finished submitting my form at 7:12, all I got was a waiting list. But that came through the next morning.

Housekeeping Notes
My interview is now live.

Sunday, February 9, 2014

DNA says We Are Fourth Cousins (Part 2 of 2)

Part one of this piece appeared last week, with the first of two examples. I am repeating the opening paragraphs here.

Bennett Greenspan, of Family Tree DNA, opened his presentation at the IAJGS Conference in Boston with something like "If we just looked at the raw data, all those of Ashkenazic Jewish descent would show up as second cousins."
This, of course, is a result of the relatively small pool of Jews in Europe a thousand years ago and the fact that for hundreds of years, most of them married within that pool. So your actual sixth and seventh cousins may look several generations closer because you have a handful of connections on your "other sides" two-three-four or more hundred years ago.

So Bennett says that FTDNA has an algorithm for its Family Finder (autosomal) test that takes this into account in order to give a more realistic estimation of cousinhood.



But even after that adjustment, I show 342 matches that FTDNA says are "suggested fourth cousins." That's not 342 fourth cousins in the world, that's 342 fourth cousins who have done Family Finder testing with FTDNA.  A result like that is patently absurd.



For comparison, they show me with forty-eight suggested third cousins.



The question is, are any of the suggested fourth cousins truly that. Let me bring two examples.
Now for the second example.
        
Dalia and me
I had long believed that Dalia was my fourth cousin. I'll spare you the long explanation, but my theory was this.
It appeared to me that the naming patterns indicated that my great-grandfather Hersch Pikholz had gotten his surname from his mother Rivka Feige and that she was a sister of Dalia's great-great-grandfather Mordecai. (Actually, it could have been the wife Taube who was the Pikholz sister of Rivka Feige, but I am trying to keep this explanation simple.)

Rivka Feige's husband Isak Fischel's surname was a mystery to me and I spent a lot of time examining Isak Fischels from all over east Galicia, looking for some clue to his surname.

This structure would have made their sons first cousins, our grandfathers second cousins and our fathers third cousins. So Dalia and I (marked in green) would have likely been fourth cousins.
   
Back in 2011, I did a Y-67 (male line) DNA test on the off chance that some match would prove to be from Isak Fischel's family.

When we took the plunge and began our family DNA project a year later, Dalia's nephew, Zachy, did a Y-37 test in order to compare him to other Pikholz families from Skalat. (In fact, he is a perfect match to the person I called "Y" in last week's example.)

But Zachy also matched me perfectly. That meant that the connection between Dalia's family and mine was not Rivka Feige, but Isak Fischel. The naming patterns still indicated quite strongly that Rivka Feige was a Pikholz, so my great-grandfather would have had a double dose of Pikholz DNA, one from each parent.

The calculation tool that Family Tree DNA provided showed a probability of about 93% that our most recent common ancestor was the father of Isak Fischel and Mordecai. There is still some chance that they are not brothers, but rather cousins.- and if so, we don't know whether first or second or even further. The science cannot be that precise.

In the course of our project, Dalia did a Family Finder (autosomal) test, as did several members of my close family. Her results showed her to be a "suggested third cousin" to my father's sister and a "suggested second cousin" to my father's first cousin Herb. Under the best of circumstances, they cannot be that close, as both cannot be closer than her third cousins once removed. But for our purposes, these results verify the idea that Isak Fischel and Mordecai are brothers.

Dalia's match with me is suggested fourth cousin." Perfect. Just what we wanted. I asked Bennett what he thought and he said I should definitely record Dalia as my fourth cousin based on the combination of our Family Finder results and Zachy's Y match with me.

How we overlap with Dalia together: Herb-41.98 cM, A. Betty-37.09 cM,Israel-13.83 cM and Lloyd 28.86 cM

Above is the set of matches that we have with Dalia on the seventh chromosome. Herb is the orange, Aunt Betty is blue and I am green. The purple is Lloyd, a certain-but-unproven third cousin of Dalia, an apparent descendant of Mordecai and Taube. This makes it clear that the connections we have with Dalia are all from the same source.

But it isn't really. First of all, our line has that "extra dose" of Pikholz DNA from Rivka Feige. So we almost have to be getting results that are closer than they ought to be. I guess that can explain why my aunt and Herb are a bit closer to Dalia than predicted. 
      
But it still leaves a problem with my match with Dalia. She shows up as my "suggested fourth cousin,"  just like 341 other people. (That number has grown in the last week to 343.) Most of those are surely not my fourth cousins, but appear to be because there is DNA from several different directions. Why should Dalia be any different?

Is the fact that we appear to be fourth cousins actually evidence that we almost certainly aren't?

And it's even more complicated
Just for fun, I looked at the same chromosome #7 comparison for the same five people, but changing the person in the background. Here is what I got.
The block on the right with all the matches is pretty much the same, no matter who is the "background" - with one exception. With Dalia as the background, Herb, Aunt Betty, Lloyd and I have a nice overlap, all matching Dalia and matching one another.

But using the others as background, my match (the green) disappears from that block entirely. How can all four of us match Dalia at the same place on chromosome #7, but I don't match the other three. The  answer must lie in the fact that there are two of each chromosome - one that came from the father and one that came from the mother. 

These results seem to be telling us three things.
  1. I match Dalia's "other" chromosome not the one that Herb, Aunt Betty and Lloyd match..
  2. My match may have come from my mother's side.
  3. Dalia's two #7 chromosomes have that same patch, perhaps because her parents are related a few generations back.
Or maybe not. European Jews have been marrying within the tribe for generations, so there are overlaps all over the place.

But clearly some of my match with Dalia comes from someplace else, other than .the last few generations of Pikholz. We only appear to be fourth cousins. Or not.
 
It gives me a headache.

I definitely have to learn more and I'm working on that.
    
Housekeeping notes
1. .I have set up a new twitter account only for genealogy. Follow at @allmy4parents.  The old Twitter has been renamed @IsraelP_Jlem and is for non-genealogy matters..

2. My interview on The Genealogy Professional Podcast goes up sometime Monday. 

3. Sunday, the day this  goes up, is the yahrzeit for my father's father. I had just turned nine when he died  and was not invited to the funeral. I never actually asked to go, but I really wanted to. I was unhappy with my parents for years about that. The first time I even saw his grave was seven years later when his brother was laid beside him.

I told my mother years later that I thought they had been wrong about this and she said that there were no children there (I am the eldest grandchild) and that some of the older relatives would have objected had I come. 

I knew who she meant.
 

Sunday, February 2, 2014

DNA Says We Are Fourth Cousins

Bennett Greenspan, of Family Tree DNA, opened his presentation at the IAJGS Conference in Boston with something like "If we just looked at the raw data, all those of Ashkenazic Jewish descent would show up as second cousins."

This, of course, is a result of the relatively small pool of Jews in Europe a thousand years ago and the fact that for hundreds of years, most of them married within that pool. So your actual sixth and seventh cousins may look several generations closer because you have a handful of connections on your "other sides" two-three-four or more hundred years ago.

So Bennett says that FTDNA has an algorithm for its Family Finder (autosomal) test that takes this into account in order to give a more realistic estimation of cousinhood.

But even after that adjustment, I show 342 matches that FTDNA says are "suggested fourth cousins." That's not 342 fourth cousins in the world, that's 342 fourth cousins who have done Family Finder testing with FTDNA.  A result like that is patently absurd.

For comparison, they show me with forty-eight suggested third cousins.

The question is, are any of the suggested fourth cousins truly that. Let me bring two examples.

Is Moshe Hersch the son of Nachman?
I have long believed that Moshe Hersch, the top of the RITA family is the son of Nachman (b. ~1795) of the LAOR family. I believed this because these were two of the only three Skalat Pikholz families where the given names Nachman and Getzel put in an appearance. (The third is the TONKA family, which I discussed two weeks ago.)

So when we began our family DNA project, it seemed to me that this was one of the theories we would actually be able to test.
A & B are second cousins. X & Y are probably half-first-cousins-once-removed, but it's a bit more complicated..



Two people from each of the two families did Family Finder tests and because of their requests for privacy, I refer to them here as A, B, X, and Y.  A and B are female second cousins from the RITA family, but actually they are a bit closer genetically because A's grandparents are Pikholz first cousins.  

X and Y are male descendants of Josef Pikholz, of the LAOR family. X is his great-grandson. Y is either the grandson or the great-grandson of this same Josef Pikholz, but from a different (non-Jewish) woman. The diagram above shows Y as Josef's grandson.

(Y has also done a Y-37 and matches other Pikholz perfectly, so we are quite certain the family story about Josef is correct.)

If Moshe Hersch is Nachman's son, A and B are fourth cousins of X,  and third or fourth cousins of Y, so that is what I was looking for. But that was before I realized that a designation by FTDNA as a "suggested fourth cousin" is likely a generation or three more distant..

Here on the left are the results as suggested by FTDNA. B is correctly matched to A and X, but not matched with Y at all. Y is further than expected from everyone, probably because his other side is not Jewish. A and X show up as third cousins instead of fourth, but that can be explained by A's double dose of Pikholz DNA due to the first cousin grandparents.

But as I wrote above, if FTDNA gives "fourth cousin" results that are not really fourth cousins, can we really accept these results when they are convenient? That doesn't sound rigorous to me.

So I did two chromosome browsers, one based on X and one based on A. (I couldn't do browsers based on B or Y because they do not match one another.)
The matches shown here are three or more centiMorgans.


























Nowhere on these two analyses do we see a place where all four match.

Both analyses show a tiny overlap of A, B and X towards the left end of chromosome 8. There are also overlaps of B, X and Y on chromosome 12 and of A, B and Y on chromosome 5.

I'm not sure that implies the kind of relationship we really want, but maybe it is.{shrug}

This is longer than I had anticipated, so the second example will wait for next week.

Housekeeping notes

1. We are now into Adar and next Sunday we remember my father's father

2. My friend and fellow Skalat researcher Jurek Hirschberg is now an authorized translator between Polish and Swedish (both ways). And he speaks a fine English.