Thursday, August 24, 2017

Progress - The Problem Is Acknowledged

Nearly six months ago, I wrote in this space about a problem at Family Tree DNA that had been bothering me for a year or more. I added a few more bits of evidence later that same day.

The problem, which I first noticed in the "Close Match Alerts" they they send every two or three days, was that people who tested more recently were being called as better matches than people who had tested earlier, despite what their own numbers were saying.

Here are three matches between a man named Mark and three of my siblings. My brother, whose test was received in Houston on 7 January 2017, matches Mark with 143 cM total and a longest segment of 8 cM. FTDNA calls him a "second cousin - fourth cousin."

My sister Sarajoy, who tested in July 2014, matches the same Mark with 145 cM total and a longest segment of 8 cM. FTDNA calls them "Fifth cousin - Remote Cousin."

My sister Jean, who tested in December 2014, matches Mark with 149 cM with a longest segment of 23 cM. FTDNA calls them "Third Cousin - Fifth Cousin."

Something is clearly wrong. It is as though the algorithm was changed but only for people who tested more recently.

If you look at the comments in my March blog, you will see that Lara Diamond - a serious DNA researcher by any standard - had noticed the same phenomenon.

I had some half-hearted correspondence with FTDNA's Help Desk, but it was all dismissed as the randomness of DNA. Then last winter I sat with Janine Cloud at RootsTech and got more specific. She sent me back to the Help Desk and there it remained for the last six months.

At the IAJGS Conference in Orlando, I sat with Bennett Greenspan - the Man Himself - and he asked me to forward my anecdotal data to him personally. Yesterday he acknowledged that the problem is real and that they would be addressing it, though not immediately. I'll mark it for follow-up in sixty days.

In the meanime, my second cousin Marty's test arrived in Houston this week and I'll be looking to see if his matches are significantly different from those of his sister Rhoda, who tested way back at the beginning.

Now if we could get GEDmatch to acknowledge their problems, we would be getting somewhere.

Housekeeping Notes
I am speaking on Thursday 7 Spetember for IGRA in Modiin at seven thirty. The address is Yigal Yadin 41, ground floor and I'll be speaking (in English) on
Lessons in Jewish DNA – One Man’s Successes and What He Learned On the Journey

Sunday, August 20, 2017

Why Do We Test? - the JewishGen Version

JewishGen is a fine organization which has been working for the last thirty years serving - even forging - the community of genealogy researchers with Jewish ancestors. As they themselves tell us, they are "[p]rimarily driven by volunteers" which often means an uneven level of expertise.

Last week, a woman whom I shall call R, posted the following question on the DNA Testing Discussion List.
"Is there any reason to do or not to do the Big Y test?  What would we learn? "
I was taught to begin answering with brevity and only when pressed, go into detail. I replied:
If we need to know what we expect to learn before testing, we wouldn't test
much and we'd learn even less.

So much of my own progress has come from tests that my project members did before I had any idea what results were expected. This is especially the case when you consider that the database of tests has been growing by leaps and bounds in the last few years. You test now and three or five or a dozen years later some new test helps you solve some problem you didn't even know to describe when you tested. And when it comes to testing the older generations, the rule is always DO IT NOW. Even if you haven't a clue what you are after.

So R asked her question on the discussion list and I replied.

The anonymous "Moderator on Duty"chimed in as follows:
Your comment, which you should feel free to send privately to R if you're so inclined, would likely be read as disparaging her question, even if you didn't mean it that way. We will not be posting it.
For the record, we all know what to expect most of the time, and it doesn't stop us from living our lives. In the world of scientific research, grant applications generally require a statement of what the applicant expects to find. Statistical comparisons require a null hypothesis, i.e., an expected result, against which to measure the experimental results. And even infants know what to expect; they react more when they encounter something different from their previous experience. So R's question was a perfectly reasonable one.
A few minutes later, the JewishGen Helpdesk sent the following:
Your issue has been received by JewishGen and assigned #10995. A volunteer will follow-up with you as soon as possible.
You can view this request's progress online here:
I am guessing this had to do with the above exchange, but I have no idea and I never requested intervention. The link they suggested didn't work.They have acknowledged my follow up inquiry, but they still haven't told me what the "inquiry" was about.
As it happens, Blaine Bettinger - one of the top genetic genealogists - had just initiated a discussion on Facebook a couple of days earlier which included this:
I've noticed an interesting misconception among genealogists when it comes to DNA; namely that we must approach DNA testing or our DNA test results with a research question in mind. But that misses out on 95% of the fun of DNA!
The response to Blaine's post was overwhelmingly in favor of testing for its own sake, without knowing what will come of it.

I don't know who the anonymous moderator is. I do know that when the discussion is about opinions, research strategies and other non-fact-based matters, the moderators have to understand that there are many legitimate approaches.

Oh, and I did write to R directly. She had no problem with any of what I said in my original response.

Housekeeping notes
I am waiting for a time and an address for a talk I have been asked to give in Modiin (in English) on 7 September.

Tuesday, August 15, 2017

Tara and Cousin Herb

Tara's matches
It began last Thursday mornng with a note from professed newbie Tara in Toronto. It seems that when she sorted her GEDmatch results by "longest segment," her top four matches all pointed to my email address.

The first of the four, with a longest segment of 70.7 cM is my father's late first cousin Herb. The other three are my sister Amy, my brother Dan and I. Worth a look? I'd say so. I mean who has a longest segment of 70 cM with a stranger?

Tara matches forty-one members of my project, but it is an eclectic group. All six of my parents' children were there, but not my father's sister and brother. And despite my X match, which is from my mother's side, there is only one second cousin from that side.

I did a few chromosome browsers and distilled the results to a single chromosome of interest - chromosome 3.

The small bit on the left side of the centromere is something I wouldn't bother mentioning, but it's right here in front of me. The five blue segments are my parents' children and the 12.5 cM green segment is my half second cousin Fred. The triangulation is good. So while this is not a large match, it does point someplace fairly specific - my Rosenzweig/Zelinka side from Trencin County Slovakia. That does not fit Tara's surnames or geography, which is largely Romanian.

The action is on the other side, where we begin with Herb's 70 cM segment in orange. On the left side of that segment, my sister Amy has 39 cM. On the right side of Herb's segment on lines 5 and 7, my sister Judith and my second cousin Terry have ~23.5 cM, my sister Sarajoy (line 6) has nearly 20 cM and the mysterious Vladimir (line 8) has 13.5 cM. On lines 3 and 4, my brother and I have ~36.5 cM that begins overlapping with Amy's segment then runs to the far right end like the others on the right. Finally my second cousin Rhoda (line 9) has a 9 cM segment that falls under Amy's and a 23.7 cM segment like mine and Dan's but not as long.

Tara's Romanian geography does not fit our eastern Galicia, nor does she have any of the very few surnames we know on the sides we share with Herb.

Of course, the thing is that if Herb has a legitimate 70 cM match with Tara, I would guess that the MRCA would not be expected to be much before 1800.

Herb's chromosome 3 on the segment that matches Tara
Before anything else, I did a triangulation to see whether all eight of the Pikholz descendants actually match Tara and Herb in the same places - or whether perhaps Herb and Co. match Tara on one of her parents and some of us match Tara on her other parent. I did that by comparing all eight (plus Tara) to Herb. Tara is on line 4 and the black vertical lines mark the ends of her match with Herb. That is the area we are interested in for triangulation.

Amy is on the left. Judith, Terry and Sarajoy are on the right. Rhoda, Dan and I overlap both sides. This is the same as our match with Tara above, so the triangulation works. Vladimir matches Herb, but not here, so he is no longer part of this analysis. He does match Tara, but the common ancestor is through her other parent.

Matching Segments
My next move was to look at Tara's Matching Segments analysis to see who else matches her long segment with Herb. (I have referred to Matching Segments before. It is a GEDmatch tool on Tier1 which requires a small donation.) I decided to limit this search to segments of 13 cM or more, but I was really interested only in those over 15 cM.

These are the results with personal information redacted and my family in red. (In this case, there is no significance to the colors on the bar graph.)

I added in the small segment from Rhoda. Eliezer is Judith's son and Avi is Judith's twin's son, so I had excluded them from the chromosome browsers above.

I also added in my fourth cousins, the half-siblings David and Anna (the second and third from the top) even though they are only about 11 cM. They did not show up on Tara's chromosome browser but they did show up here. Same thing with my third cousin Joe. In all three cases they show up on a one-on-one with Tara, with a single matching segment. (GEDmatch probably has some obscure rule that explains this.)

All three triangulate with Herb.

What is significant with all three is that they are not descendants of my great-grandmother, so Tara's segment with Herb and Co. appears to be a Pikholz match, not a Kwoczka match. (There is a reason I say "appears to be" but I won't go into that here.)

Just for fun, I ran Tara on the fancy one-to-many at Tier1 and the three single-segment matches don't show up there either. Not even when I raise the limit from 2000 to 5000. Furthermore, the 2000-match Tier1 search does not give all the normal matches that the regular 2000-match search shows. Much as I love GEDmatch, sometimes you have to look at something half-a-dozen ways to get a full picture and then hope you got it right.

Triangulating Tara's non-Pikholz matches with Herb
I did one-to-ones between Herb and each of the twelve people in the above list whose matches with Tara are 15 cM or more. Leah and the three in the "Elisha group" do not match Herb at all, so they are on Tara's other side.

E, J, Carol, Steven and Ryan match more or less according to their matches with Tara, though Ryan's is about 30% larger.

The other three - L, Harriett and Joseph - match Herb on no more than 10.5 cM of their matches with Tara. What about the rest? Who knows. It will require further inquiry.

What happens next?
Before I draw any conclusions and even before I write to any of these matches, Tara's mother has agreed to test. Probably her father too. That should clarify whether Tara's long match with Herb is from one of her parents or is perhaps - as is likely - a composite, part from her father, part from her mother. The "Are Your Parents Related" tool on GEDmatch shows her parents unrelated to one another on chromosome 3, so we needn't worry about that.

Oh, and Tara has a paternal grandmother who is willing to test. Anyone in Montreal who can help with this?

Meantime, Tara has joined our surname project at FTDNA.

AMENDED- Follow up posts here and here.

Monday, August 14, 2017


My mother's mother, Sarah Rosenbloom Gordon, went to the US from Borisov before WWI - as did her two younger sisters and her younger brother. The two sisters were dead by early 1916. An older sister stayed in Russia with her husband and children. There was also a younger half-brother and half-sister who remained in Russia with the father and step-mother. The half-brother was killed in a pogrom of some sort after WWI.

Less than seven moths ago, I reported here that we (=Galit Aviv-Sisto) had found the grandchildren of the older sister Alta Rosenbloom Kaplan. Grandsons in Indiana and New Jersey, a granddaughter in Columbus Ohio, a granddaughter in Nuremberg, a grandson who died here in Israel (I am still looking for his three children) and at least two living granddaughters in Moscow.

The IAJGS Conference in Orlando was three weeks ago (seems more like three months!) and Aeroflot has a nice price from Israel to the US east coast (< $1050), so I booked my flight and arranged to visit one of the Moscow cousins - my second cousin Liya, her husband Gennady and their daughter Katya who has been my Facebook translator these past months. The other, older cousin was not well and her daughter who would have accompanied us, was out of town.

My three "new" second cousins - first cousins to one another - in three countries. I now have DNA from all three.

Katya met me at the airport Tuesday evening. Moscow is a huge city, very sprawling - at least that was my impression - and it took about an hour to get to their apartment. The weather was much warmer than I expected. Liya is a retired mathematician, as was her father (my mother's first cousin) and his younger sister. Gennady is a retired professor of electrical engineering.

They have a son, Ilya, who lives in UK but as his daughter Tatiana was already in high school when they went, she stayed behind to finish her schooling in Moscow. Tatiana (twenty-one) is studying mathematics. When I told her I was going to Orlando and identified it with Disney World, she said that all she knows about the US is where people study mathematics. These mathematicians tend to be focused.

Wednesday, we went to the cemeteries, Gennady driving, Liya with her gardening tools, Katya and I. Our Kaplans are all buried in the Vostrikovsky Cemetery. In the best Soviet tradition, it is non-sectarian and perhaps that is the reason it is in such good condition. There is no Hebrew on the tombstones, nor are there any Jewish symbols. Not even the small stones that visitors leave in many countries. Each grave or pair of graves is surrounded by a metal fence.

My great aunt Alta is here as are all of her children. It is not clear where her husband is buried or even when he died, in the early 1920s. Apparently he was never discussed. Uncle Hymen said his name was Berl, but his children are Borisovitch and Borisovna. You will see the problem with that further on.

Most of our Kaplan graves are in section 94.

Next to the road is the grave of Sophia (Sonia) Borisovna Kaplan, born 1921, died 1969. I do not have precise dates. She was a mathematician and died of breast cancer.

Her son David (who is almost certainly named for our great-grandfather, as am I) is the one who came to Israel. Next to her is her husband Maksim Markovitch Gorohov, born 18 April 1925, died 30 April 2004.

Maksim's surname was originally Razgon but was "sovietized" to Gorohov. David reverted to Razgon in Israel.

This is the view from the side of Sonia's tombstone.  Sonia shares the grave with her older sister Esfir Borisovna Kaplan, 1903-1978.

Esfir - which I am told is the Russian form of Esther - was named Etta Bryna after her maternal grandmother who died in Borisov in 1896, probably before her fortieth birthday. Etta Bryna's surname is unknown and I have written abut this several times, most recently here.

Esfir's husband was Yosef Gemeyner. She was not able to have children but she raised her husband's son Mark from his previous marriage.

Alta, my grandmother Sarah and Uncle Hymen all had daughters named Etta Bryna for their mother and all three of those daughters adopted after not being able to have children. (My Aunt Ethel had two natural children after adopting.)

A bit further back from the road, without moving the car, are Liya's parents.

Isaak Borisovitch Kaplan 1909-1966 and Emma Abramovna Elkind 1913-1969.

Liya, their only child, ws born in 1939, so she was orphaned of both parents by age thirty. Alta was still alive when her youngest children, Isaak and Sonia, died.

The picture of Isaak on the tombstone - a Russian custom which is practiced in Israel as well - I have never gotten used to it! - is certainly familiar.
This is the same Isaak, standing on the right. This was the only Kaplan photograph we had until seven months ago, so we thought there were only four children, not six. (The girl is Sonia.)

Below, we see the Kaplan family in 1914, just before Uncle Hymen (standing in the back) sailed to the US on the last ship before WWI. Isaak is the young boy seated on the left. The daughter on the table died and I do not know her name. Esfir is the girl on the right.
A bit further back at the same stop are the graves of Lev Borisovitch Kaplan, born 18 March 1907 and died 4 Decenber 1986. His wife Pelageya Alekseyevna was born 13 November 1912 and died 23 January? 1989.

I assume Lev was named for Etta Bryna's father Yehudah. We have no idea when or where he died, but there is a good chance it was after Etta Bryna.

They had three children - one now in new Jersey, one in Columbus and the youngest who died in Moscow ten years ago.

As we visited each set of graves, Liya brought out her gardening tools and went over each one. The clippings were carefully removed.

We moved along the road to the corner of the section where we found the grave of Boris Borisovitch Kaplan, 1900-1973. This is where we defer to Uncle Hymen's memory of Alta's husband as Berl, as there is no way Boris would have been given the same name as his father. In fact, Uncle Hymen said that this son, Boris, was named Boruch Yosef.

His wife Genia Yankelevna Chankina - 1897-1972 - is in the same grave. To the right is a sign - not a proper stone - for her brother.

The fence was freshly painted. Perhaps at the direction of their daughter Dusya. Or Dusya's daughter. Dusya was born in 1929 and is one of the two great-granddaughters mentioned by my great-grandfather Israel David Rosenbloom in a 1929 letter to Uncle Hymen. No one is certain if Dusya is alive as the families are not in contact.

We crossed the main road to the other side of the same cemetery heading for section 116. On the way we encountered an entourage of large men in suits, sunglasses and Mercedes looking for all the world like your stereotypical Russian mafia, straight out of central casting. We carefully found a way around them.

The eldest son Yakov, named for Israel David's father, is in section 116. Yakov was a shoemaker which enabled him and all the family to live in Moscow. This is what saved them during the Holocaust.

Yaakov is not called Borisovitch, but Berkovitch which fits better with Berl as Alta's husband's first name. My mother's first cousin, Yakov Berkovitch Kaplan was born 1899 and died 1984.

His wife Fanya Gershevna Pinskaya is buried in the same grave. She was born 1905 and died 1988. The Soviet regime did not acknowledge marriages at that time and their children have Fanya's Pinsky name.

Fanya was from Penza, where Israel David lived after WWI, and I assume that is how Yakov and Fanya met. Inna in Nuremberg is their daughter.

Within the boundaries of the same fence is Alta.
Alta Davidovna (from Israel David) Kaplan, 1879-1970. The family tradition was that she lived to ninety-three, so the stone may be incorrect. Or the tradition.

The family says that the name "Alta" was added during a childhood illness and that her original name may have been Masha - but no one is sure.

There are about six years between Alta and my grandmother Sarah, the second child. That fits with the story that Uncle Hymen was called CHAIM Benzion because several sons died in infancy. Very likely they were born between Alta and Sarah. (Attention family members - Benzion and Masha are probably family names. Chaim and Alta are not.)

Finally - on this day which was beautiful for visiting cemeteries but miserably hot for sitting in traffic - we drove to the Malakhovskoye Cemetery to see Aunt Mera, the half sister with whom Uncle Hymen was particularly close.

The stone on the left is Maks Yakovlevitch Goldin 1905-1978 and Mera Davidovna Goldina 1903-1990. My grandmother, who died in 1959 in her seventies, had a living sister as recently as twenty-seven years ago. Aunt Mera was a physician and the cemetery records say she was from Penza.

I do not know if here is any symbolism associated with the cut corner on the top left of the tombstone.

This is a Jewish section, rather than standard Soviet, and people leave stones. The flowers are plastic, so could have been there for years.

The grave on the right is Mikhail Maksimovitch Goldin 1931-2009. Aunt Mera's son. The blank space below his name looks like it should be for his wife, so she may well be alive. Maybe children too.

At four in the morning, it's already light that time of year in Moscow. Gennady and Katya drove me to the airport at five thirty Thursday morning for my 9:25 flight to Dulles. There were eight chairs at the gate.

The only people I saw during my Baltimore-area overnight were Uncle Bob and Ro and Friday I was off to Orlando.

Sunday, August 6, 2017

The Wives of Moshe Hersch

Mail from Judy
In Orlando with Rozdol Pikholz cousin Audrey (left) and
my fourth cousin once removed Steve and his wife Diane
I have been very busy lately and have not even written yet about my visit to Moscow. And I'll not be writing about the IAJGS Conference in Orlando.

But some DNA matches came up this week that appear significant and I want to tell you about one of those. (Despite everything, I have been trying to keep up with all these "I have ## matches with your project. How are we related?" emails.)

This one (partially redacted) came Thursday:
Dear Israel Pikholz,
Your reputation precedes you. However, I didn't think I had many matches with you until I looked at it from the perspective of my (full) brother's DNA report, and especially from Chromosome 23. I am attaching a PDF showing all 17 matches of individuals you manage who appear on my brother's list in GEDmatch.

My brother is Ira Rosenberg ... and I am Judy Pritchett .... I have included myself.

Our four grandparents' surnames are:

Rosenberg (unknown, possibly Zbaroz) ff
Dejur/Deshur (Katerinopol), fm
Abramowitz (Odessa), mf
Alter/Alpert (Kriulyani, Bessarabia) mm

On Chromosome 23 there do seem to be two distinct family lines represented. 
This is chromosome 23, as she attached it.

Line 1 is Judy herself, where she matches Ira.

Lines 2 and 3 are my fourth cousin Nan and her son Steve who introduced one of my Orlando talks, ~25.6 cM. Line 4 is Nan's third cousin and my fourth cousin Charlie, ~16.3 cM. The four blue bits (5.3-5.9 cM) are Aunt Betty, two of my sisters and Uncle Bob's daughter. When I ran it myself the following day, Uncle Bob himself also appeared, as did my double second cousin Marshal. (Marshal was missing because Judy only saw seventeen matches with my families, when in fact there are forty-nine.)

All these matches are on my father's side - and in fact must be my grandmother's side because Uncle Bob gets no X from his father.

The large X-match between Ira and Nan & Charlie

Nan and Charlie are, as I said, my fourth cousins, descendants of Moshe Hersch, the brother of my great-great-grandmother Rivka Feige Pikholz.

Charlie could have received the segment in question from either parent of his great-grandmother Sara.

But Nan could not have. Her great-grandfather Berisch received X only from his mother. Since she and Charlie received the X from the same source, it must have been the mother of Berisch and Sara - Moshe Hersch's wife. (I acknowledge that it is possible that the X in question snuck into the lineages of Charlie and Nan via a spouse of one of the intervening generations, but based on what I know of the family, I consider this highly unlikely.)

Ira's Rosenberg family comes from Zbarazh, which is near Skalat, but since Ira received no X from his father, that cannot be the source of this match. But that is another question, not for here.

But now I must introduce a bit of research history. 

The Hebrew side of Berisch's stone.
When we received the 1826-1845 Skalat deaths maybe eighteen years ago (thanks to Jacob Laor and Alex Dunai), we found an 1843 death for seven year old Israel Pikholz, son of Moshe and an 1842 death for three year old Perl Pikholz the daughter of Moshe Hersch. The fathers may be the same person, but not necessarily. We also have the 1918 death of Nan's great-grandfather Bernard / Berisch Pickholtz in Philadelphia - which includes his birth date 14 December 1837. His father is named on his tombstone as Moshe Zvi, which is, of course, Moshe Hersch.

So it appears that Moshe Hersch may have had these three children in a three year period, before 1840.

Later we received an 1887 death record from the AGAD archives for Sara Pikholz, showing her to be forty years old. From the birth records of Sara's children, we know her parents to be Moshe Hersch and Jente. (Sara's husband is Szulim Pikholz, but that has nothing to do with this analysis.)

Sara's 1887 death record - age at death was forty.

Because of the long break between the three pre-1840 births and the 1847 birth of Sara, I was not prepared to declare that this was one family. When I got into DNA, I demonstrated in Chapter Six of my book that the two Moshe Hersch were the same person, but I was still wary of saying that all four children were from the same wife. (Yes, I know that there may have been additional childen in between, but I have seen nothing that mentions such children.)

But since Nan and Charlie share a segment on the X chromosome that can only have come from Moshe Hersch's wife, it is tempting to put the question to bed by declaring that Moshe Hersch had only one wife - the one we know as Jente.

But not so fast. If The mother of Moshe Hersch's pre-1840 children died, it is likely that the family would have married him off to someone else within the family. The dead wife's sister or niece, for instance. So even if Moshe Hersch had two wives, there is a better than even chance that they would have had much of the same X chromosomes. If they were sisters, the X from their father would have been identical.

I still do not know if Moshe Hersch had two wives, and we may never know. But I am now certain that if he did, the two women were closely related. Considering all the difficlties in this research, that qualifies as progress.

What about Ira's matches with my grandmother?
In addition to the six ~5.5 cM segments, Ira also has matches of more than 10 cM with my grandmother's family, on two other chromosomes.

On chromosome 17, Ira has matches of 13-16 cM with Aunt Betty, Uncle Bob, my second cousins Marshal and Susan and one of my sisters. Another sister has ~8.8 cM on the same segment. So we know that Ira is related to my grandmother's side, as I showed above on chromosome 23.

On chromosome 18, there are three identical segments of ~15.9 cM with my father's second cousin Shabtai, one of my sisters and me, plus a segment of ~12.6 cM with another of my sisters. This too is my grandmother's side, but in this case Shabtai makes it clear that the match is on my grandmother's mother's (Hungarian) side. The truth is, I'd be more comfortable if Aunt Betty, Uncle Bob or one or two of the second cousins also showed up on this segment, but we can only use what we have.

This match on my great-grandmother's side does not in any way indicate that the other two segments (on chromosomes 17 and 23) are also my grandmother's mother's side, rather than her father's side.

Now if Judy and Ira have some surnames and geography that fit any of this, or if they have cousins who can test, we may be able to move this along.