Tuesday, June 27, 2017

First Cousins / Genetic Half Brothers

My nephew Eliezer's Family Finder results just came in. That's one of the sons of my sister Judith. Her twin - my sister Carol/Devorah - was killed in an auto accident thirty years ago, but since they are identical, we have her DNA. Except for the astigmatism and later the glasses, you could not tell them apart.

But Judith wanted to know what the DNA would say about their children so we ordered Family Finder tests for Eliezer and for one of Carol/Devorah's sons Avi. Avi's results came in a couple of weeks ago and I discussed them here.

This is the chromosome browser for Judith showing the two boys. There is no difference between the orange and the blue and you cannot tell which is her son and whch is her nephew.

Family Tree DNA says each of the boys shares 3382 centiMorgans with Judith, with a longest segment of 267 cM.

The two boys match me quite closely (Avi 1788 cM and Eliezer 1781 cM) and the others not quite so closely - Dan (1771 / 1837), Jean (1793 / 1705), Sarajoy (1635 / 1684). But they differ quite a bit regarding Amy.

Amy and Avi share 1495 cM with a longest segment of 83 cM. Amy and Eliezer share 1718 cM and a longest segment of 113 cM. This fairly large difference doesn't mean anything - it's just the randomness of recombination.

There is, however, one number worth noting. Avi and Eliezer share 1963 cM, based on the FTDNA match page. On GEDmatch, it's even greater, with 2177cM at the standard threshhold of 7 cM, 2223 cM with a threshhold of 1 cM.

The International Society of Genetic Genealogy Wiki says that half siblings share about 1700 cM, so our numbers appear high. Blaine Bettinger's Shared cM Project brings actual statistics showing that half siblings share anywhere from 1320 to 2134 cM with an average of 1753 cM. (His statistics show full siblings to be 2150-3070 cM with an average of about 2600 cM.) He calls the range 2100-2230 cM the "caution area," where you cannot determine the relationship from DNA alone - maybe half siblings, maybe full siblings.

It is tempting to say that the extra-large overall match between Eliezer and Avi may have something to do with the fact that their fathers' families come from the same general area - and both are kohanim, to boot - but no one has an interest in actually checking this out. I showed some of these numbers to Blaine and asked if it is reasonable to say that "there is a good chance their fathers are related" or perhaps "a very good chance." He allowed for the possibility that the fathers are related but "probably not closely." I can go with that.

Sunday, June 18, 2017

Success Using Family Finder Match Alerts


Family Tree DNA Alerts
I manage over ninety family kits, most of whom tested with Family Tree DNA. So I get many scores of these notices every week. Not only each person, but since, for instance,  Dan is a member of all five of my projects, I get all of his notices five times. Others are members of multiple projects and for them too I get multiple notices.

These notices of "Close Matches" means matches that FTDNA considers to be suggested second-third cousins. Readers may recall that several months ago I challenged FTDNA to explain why the more recent kits are getting far and away more close matches than earlier kits from the same families. We are four months and counting and FTDNA has not yet addressed this issue. (Janine, if you are reading this, see my help request 565856  which was opened 17 February.)

I cannot just limit myself to those matches which FTDNA calls "close." I also want the 2-4 cousins and the 3-5 cousins. After all, if one person has is a suggested 2-3 cousin, it is relevant to consider siblings or other close relatives who are 2-4 or 3-5 with that same person.

So between the huge numbers of alerts and the inconsistency in reporting them, I have yet to find a way to make use of them properly. After all, if I get a close match with Joe Schmoe, I cannot look at his matches with my other kits. It all becomes unwieldy and my worry that I will miss important matches competes for my time with more immediate demands.

An Attempt to Manage the Alerts
A few weeks ago, I decided to try something new. I decided to download in an Excel file all the matches during the month of May for each of my project members. then I arrange them in separate Excel files by group: my mother's side, my grandmother's side, the Rozdol-Pikholz side, etc. That's a lot of work by itself but after that I have to sort by the names of the new matches to see who may have interesting matches with several people within each of my groups. There is no macro that will do that for me.

I then sent out over two hundred emails like this.

Some replied, most did not. Some gave me their GEDmatch numbers, others did not care to share this secret information with me. Others needed help even creating GEDmatch numbers. Oh, and a few would send me a list of all the GEDmatch kits in their families.

I looked at each one against all my kits - after sorting on the "Name" column of their match lists so all mine would come up together near the top  - and created 2-D Chromosome Browsers. For most I would do two or three Chromosome Browsers for different parts of my families.

In one case after another - particularly within the Skalat Pikholz families - I would get results that I couldn't do anything with. The only segments over 10 cM were individuals, not groups. And when they were groups, they were vague and appeared weak and distant. I mean, if I have a segment shared only by a third cousin here and a fourth cousin there and a double fourth cousin another way, how serious can this be. It almost has to be long ago and in most of my directions I have only two or three ancestral surnames to work with, even when I can go back two hundred years or more.

I was also hampered by the total inadequacy of the Tag Groups that GEDmatch inaugurated a few months ago. I have been meaning to write about that and will try to do so soon.

I really began wondering what was the point of all this work. After all, if I were serious, I'd have to do this every month! I would send the results to the matches and began concluding with "Thank you for humoring me."

First Partial Success
Last week I saw some progress. Kind of.

I heard from Ellen, the wife of one of my new matches of interest, a man named Robert. She gave me his GEDmatch number and I went to work. My Chromosome Browser gave me this:


Identical segments with two of my sisters and my brother, a similar segment with my half-second cousin Fred, and a smaller segment in the same place with my second cousin Susan. This is not large but it is unambiguous. My father's mother had a half sister (same father, different mothers) named Ella. Aunt Ella's husband was not Jewish, nor was the wife of their son. So my half second cousin Fred has all his Jewish DNA from one grandparent, Aunt Ella. Susan is a full second cousin on that side. There is no way that our common ancestor with Robert is not an ancestor of my great-grandfather - either a Rosenzweig or a Zelinka. Both families lived in the area of Trencin County Slovakia back into the 1700s.

It reminded me of the match with Cousin Debbie last year, on a segment that looked like this:









True, Debbie's segment with us is larger than Robert's and she has more matches, but nonetheless this is the same logic and I can accept Robert as a family member with the same authority.

And Robert has another match, this one with Fred and my double second cousin Lee. It is possible, though unlikely, that this comes from  different common ancestor that the match on chromosome 7, but even if so, it does not challenge the conclusion.

However, whereas Debbie is definitely Zelinka, not Rosenzweig and she knows of Trencin County ancestors, Robert's position is less well-defined. He could be either Rosenzweig or Zelinka and in any case, he knows his family to be from Horodenka in southeastern Galicia. So we have work to do here, but we know there is at least a small pot of gold to be claimed.

The Duncans
When it rains, sometimes it pours. Or at least rains a little more. The next GEDmatch I looked at after Robert was a brother and sister pair, Evelyn and Adam. Their father is Scottish, a Duncan, so my families' matches with them are on their mother's side.

I started off with this excellent set of matches for Evelyn on Chromosome 12. Regular readers will recognize them easily enough.
Evelyn's matches with my mother's mother's Rosenblooms, from Borisov in Belarus.

















  • The first two, Inna and Lydia are granddaughters of my grandmother's sister Alta. They are first cousins.
  • The next two, Beverly and Sam are grandchildren of my grandmother's brother Hymen. They are siblings.
  • My sisters Amy and Sarajoy and I are on lines 5, 6 and 9.
  • My first cousins Kay and Leonard round out the group.
Evelyn has another match with Inna and Lydia on Chromosome 20, of about 11 cM.

Evelyn's brother Adam has much the same segment on Chromosome 12.













It's a bit different from Evelyn's matches, but with the same clear message. We share a common ancestor upstream of one of our great-grandparents Israel David Rosenbloom or his wife Etta Bryna. And speaking of Etta Bryna, my maternal haplogroup, as seen in my MtDNA test is U1b1. Evelyn's is U1b. These are very close and may refer to the same common ancestor as these matching segments, though the matching segments appear more recent.

Adam has another segment that Evelyn does not.











This points in a slightly different direction. It does not have the Rosenbloom cousins, but it has five of my mother's children plus our first cousin Kay - and our second cousin on our mother's father's side, the Gordons. I am not quite sure what to make of this because Judy's Jaffe grandfather also came from Borisov. What is certain is that Evelyn and Adam are our cousins - probably fourth, maybe fifth or even third. Galit has added them to our Rosenbloom Borisov project.

The problem is, we do not know how to go from there. We do not have additional known ancestral surnames from our side and though they have a few, we cannot put it together. And their geography is Pinsk rather than Borisov - that's a distance of nearly 400 km.

I also had a look at the matches on Chromosome 12 on the GEDmatch Matching Segment tool to see if there is anyone else who matches both the Duncans and the Rosenblooms on that segment. I see none.

Now I have to decide if I want to do this again for the June matches.

Housekeeping Notes
I'll be speaking on the Hebrew version of
Lessons in Jewish DNA – One Man’s Successes and What He Learned On the Journey
this week, 19 June at 6:30, for IGS Rishon Lezion, Museum of Rishon Lezion, Ahad Ha’am 2.

Also, my son in Chicago is making his next bar mitzvah the first Sunday in May. If any program directors are looking for something around then, please drop me a note.

Wednesday, June 14, 2017

The Lydia Conundrum

Lydia
Lydia and Inna are second cousins of mine on my mother's mother's side. They are first cousins to one another. Their grandmother, Alta Rosenbloom Kaplan, is the older sister of my grandmother Sarah Rosenbloom Gordon, who remained in Russia when her sisters and brother went to the US before the First World War.

Lydia's DNA results came in two months ago and Inna's this week.

We now have autosomal DNA from thirteen Rosenbloom second cousins: Lydia and Inna from Aunt Alta, Beverly, Beth and Sam from Uncle Hymen and my first cousins Kay and Leonard, my four sisters,  my brother and me from my grandmother Sarah.

I hope to take DNA from two more of Alta's granddaughters when I visit Moscow next month.

Two of Lydia's matches with the second cousins are over 420 cM. Eight more are between 342 and 387 cM. One is 286 cM. Compare these to the ISOGG definition of second cousins - 212.5 cM - and Blaine Bettinger's Shared cM Project's average of 238 cM and we see that Lydia's matches are much higher than expected. If it were just the vagaries of DNA, some would be higher and some would be lower, but that is clearly not the case here.

So the answer must be Jewish endogamy. All Jews, being related multiple ways, have much larger matches than the general population.

Inna with her daughter, granddaughter and a visiting Israeli cousin
Inna's results are less striking. She has over 300 cM with only three of the eleven second cousins, the highest being Leonard with 386 cM. She has six more between 258 and 296 cM. Then me at 224 cM and my brother Dan at 215 cM.

This appears to be a much more normal distribution, with less influence from Jewish endogamy.

With that, Inna has between three and six matches of over 20 cM, on eight segments with the second cousin group, including matching segments of over 50 cM.

It follows that one of Lydia and Inna, daughters of brothers, has a much greater influence from endogamy than the other. And if I were to tell you that one of them has a non-Jewish mother, hence much less endogamy, you would say that it must be Inna.

You would be wrong. As I was.

Inna's mother is from a normative Ashkenazi Jewish family. FTDNA's MyOrigins calls her 93% Ashkenazi Jewish, typical of our family. There should be lots of endogamy here.

Lydia's MyOrigins shows her to be 45% Ashkenazi Jewish and 46% East Europe non-Jewish, plus some fragments. The normal sort of background endogamy is missing, so I fully expected that Lydia's large matches with us were a result of a close cousin relationship between her Rosenbloom grandmother and her Kaplan grandfather.

If that were indeed the case, Inna's matches would be even larger because she has both the supposed cousin grandparents and the standard, garden-variety Jewish endogamy. But she doesn't. So she doesn't.

Why? Beats me! More important, why does Lydia have these big numbers? What else is going on here? It looks much too large and much too skewed to be "the strange ways of DNA." (Lydia's best match with the cousins is with Kay, whose faher also has no Jewish DNA.)

I cannot wait to see what the Moscow cousins have. One of them is Inna's sister, the other a first cousin to Inna and Lydia. So unless anyone has some suggestions, I expect to revisit this at the end of August.

Housekeeping notes
I'll be speaking on the Hebrew version of
Lessons in Jewish DNA – One Man’s Successes and What He Learned On the Journey
on 19 June at 6:30, for IGS Rishon Lezion, Museum of Rishon Lezion, Ahad Ha’am 2.

Also, my son in Chicago is making his next bar mitzvah the first Sunday in May. If any program directors are looking for something around then, please drop me a note.

Monday, June 12, 2017

My Nephew Avi

My nephew Avi's Family Finder results just came in. Avi's mother - my sister Carol/Devorah - was killed in an auto accident thirty years ago. But she and Judith are identical twins so we have her DNA. Except for the astigmatism and later the glasses, you could not tell them apart.

But Judith wanted to know what the DNA would say so we ordered Family Finder tests for Avi and for one of Judith's sons Eliezer. Eliezer's was delayed in the mail - or in FTDNA's mail room - but should be along in two or three weeks.

These are Avi's matches with his mother's six siblings.
FTDNA accepts him as Judith's son, as expected.

When I looked at the six of us against one another a few months ago, I saw that Amy is not as close to Judith as the rest of us. This shows up in Avi's match with Amy as well, both in the longest segment and in the total.

Avi's match with Judith is 3382 cM.

We have five other Pikholz parent-child comparisons: Aunt Betty and her son, Uncle Bob and his daughter, my fourth cousin Nan and her son and daughter and Maxine from the Nachman Pikholz line and her daughter. Aunt Betty's son, Uncle Bob's daughter and Nan's daughter share 3384 cM with their parent. Maxine and her daughter share 3383 cM and Nan and her son share 3382 cM. All five and Avi have a longest segment of 267 cM.

On Avi's chromosome browser, I see tiny breaks in chromosomes 3 and 8, so I assume that is what is missing from the 3384. In any case, this layman doesn't see anything that implies anything special about the twins.

After I get Avi on GEDmatch and see Eliezer's results, I'll have another look.


Housekeeping notes
I'll be speaking on the Hebrew version of
Lessons in Jewish DNA – One Man’s Successes and What He Learned On the Journey
on 19 June at 6:30, for IGS Rishon Lezion, Museum of Rishon Lezion, Ahad Ha’am 2.

Sunday, June 4, 2017

Are My Parents Related

Related or not, my parents are married
The indispensible DNA-analysis website GEDmatch has a tool called "Are Your Parents Related." It's pretty straight-forward. You enter your GEDmatch kit number and it gives you a result. No options for threshholds or anything else. Just press the "submit" button.














They explain it like this:
Since you inherit half of your DNA from each of your parents, it stands to reason that large blocks of SNPs where both alleles are the same would be an indication that your parents each inherited that block from the same ancestor. These are called 'Runs of Homozygosity' (ROH). There are other utilities available that look for ROH for other purposes, but this analysis is specifically aimed at determining how closely related your parents might be.
They don't say so, but it is obvious if you think about it that the results are nothing more than an indication. After all, what they work with is your own personal DNA and since your siblings' DNA is different from yours, they will produce different results.

The GEDmatch kits of my brother and my sister Jean show that there is no indication that our parents are related. They both received the results - such as they are - on the right.

Mine was different and my sister Amy's was different from mine. Amy's kit shows that my parents share 7.1 cM on Chromosome 1. Mine shows that they share 7.6 cM on Chromosome 9.

Sarajoy's kit (below) has a third segment, with 8.3 cM on Chromosome 3.

That's 23 cM altogether.

Judith's kit (below) has two segments, Amy's from Chromosome 1 and mine from Chromosome 9.

There are six of us, so you might think that we encompass all of our parents' DNA. We probably do, but we don't necessarily have their matching segments together.

GEDmatch offers another option - one that I have written about before in other contexts, as recently as last week. It's an idea I had during the Shavuot holiday, Tuesday night.

I created a simple Lazarus kit for my father. His six children are in Group 1 and his sister and brother are in Group 2. I could have added other family members, but any of those would have introduced DNA from other sources. My father's cousin Herb, for instance, would have brought DNA from his father who in theory could be related to my mother.

This Lazarus kit is 3490.2 cM.

For my mother, I used the same Group 1, her six children, but she has no living siblings for Group 2. So here I had to use my mother's sister's daughter and her brother's son. Not quite as good and with a chance of some contamination - at least from the nephew. The niece's father converted to Judaism and would have had no DNA in common with my father.

My mother's Lazarus kit is 2823.2 cM.

I ran a "One-to-one" between my parents Lazarus kits. THAT should give me a minimum for how closely they are related. Spoiler alert, it's more than the 23 cM than we saw in our "Are Your Parents Related" runs.


Nine matching segments for a total of 95.7 cM. That's four times 23 cM. Blaine Bettinger's Shared cM Project would define that as third cousins or second cousins twice removed. My parents are certainly not related that closely. Two of those segments - on Chromosome 10 - are adjacent and they alone are nearly 23 cM.

My parents' Lazarus kits show no match on the X.

But the biggest surprise is that there is nothing on Chromosome 3 or Chromosome 9 that Sarajoy and I show. And there are two segments on Chromosome 1 but neither is the segment that Amy and Judith show. I am guessing that the Lazarus algorithm is not designed to recognize those Runs of Homozygosity and that my cousins do not have those matches with my mother.

I don't have much practical use for "Are Your Parents Related," so the discrepancy really doesn't matter to me. But it certainly is one of those things that make you go "Hmmm."

Housekeeping notes
I am posting this Sunday morning my time. This evening, I'll be speking for “Shorashim BaGalil” in Kiryat Tivon at the Library and Memorial Center Migdal Street 2. It will be the Hebrew version of
Lessons in Jewish DNA – One Man’s Successes and What He Learned On the Journey

I'll be speaking on the same topic on 19 June at 6:30, for IGS Rishon Lezion, Museum of Rishon Lezion, Ahad Ha’am 2.