Friday, 15 January 2016

Big Chunks of DNA

Autosomal DNA match: Michael Shade and CG

I am of East European Ashkenazi Jewish origins. For centuries the East European Jews lived as small minority populations in whichever country they were in, and consistently married within their own community. Married couples were often related to each other, more or less closely, sometimes as closely as first cousins. This phenomenon of near-exclusive intra-community marriage is referred to as endogamy, and it has led to a situation today in which nearly all Ashkenazi Jews seem to be related to each other, more or less distantly.

This has significant consequences for our Autosomal DNA test results. Ashkenazi Jews have many times more ‘matches’ than other groups, and the predicted relationships often seem exaggeratedly close. FamilyTreeDNA have recognised this effect, and have altered their algorithms to try to take it into account, but the results still seem skewed. This makes it particularly difficult to identify genuine matches that can be checked against evidence gained from traditional methods of family history research.

In my own case, I now have over 100 matches at a predicted level of 2nd-to-4th Cousin or closer. However, I have not been able to establish a definite relationship to a single one of them, apart from Katy, the daughter of one of my 1st Cousins. I discussed these issues in the post A Hundred Up, from a couple of months back.

Here I look at the case of a match that came up a couple of months ago with someone who, at first sight, doesn’t seem to be Ashkenazi Jewish at all.

Disclaimer: I must point out before we start that I am a total amateur at this, and may be talking complete rubbish. If you know better, or can suggest a different interpretation, please let me know. Nicely. Many thanks.

FTDNA Family Finder results
As at 14 Jan 2016, CG is my 10th closest match on FTDNA’s Family Finder. I currently have over 5300 matches there altogether, the vast majority being too remote to even consider following up. A look down any page at random suggests that the vast majority of those listed have recognisably Jewish surnames, and a check of the Ancestral Surnames that some have listed seems to confirm this.

However, of the 100-odd 2nd-to-4th Cousin matches, very few show any of my own ancestral surnames or places, and if they have one, they don’t have the other. Of the closest 20, only one lists an ancestral surname that corresponds to any of mine. So they don’t feel like a bunch of 2nd-to-4th Cousins, even if that’s how FTDNA categorises them. That’s Ashkenazi endogamy for you.

CG stands out from the crowd, as his list of 60-odd ancestral surnames includes only 3 that could be considered to be recognisably - though not exclusively - Jewish. One of his great-grandfathers was Jewish, and used the surname Lewis. His parents had immigrated to London in the 1870s, from somewhere in Poland, where their surname was Lisky.

FTDNA: Shared centiMorgans
On the FTDNA Matches page, you can sort by Relationship Range to see a listing in order of genetic closeness to you. In my case this puts my cousin Katy at the top - she’s a 1st Cousin Once-Removed - followed by a few predicted to be 2nd-to-3rds, and then the 2nd-to-4ths. CG is in 10th place - but it turns out this is not because of the total amount of DNA we share, but rather because of the length of the longest shared segment.

The unit used is centiMorgans - don’t ask me to explain what they are, I’ve seen several different explanations, none of which I can claim to understand, but for our purposes it’s the standard measure used for the amount of DNA you may or may not match someone on. The more cMs two people share, the stronger the match.

Here are the relevant numbers for CG and the four people around him in the list - these are my top 5 out of a total of 100 2nd-to-4ths:

Total shared cM
Longest shared segment, cM

So although CG has only 1 Ashkenazi Jewish great-grandparent, and so is only 1/8 AJ, he shares 1/2 as much DNA with me as the others do. And the longest segment we share is right up there with the best. Looking at these figures aroused my curiosity even more.

FTDNA: In common with
It struck me that the match with CG could be quite helpful, even if we were not able to determine exactly how and where we were related. If I could pin down on which side of my family he matched me, that might help locate some of my other matches, at least on one side or the other of my family.

This is where Katy comes in handy. I know she is on my father’s side, and I can use FTDNA’s ‘in common with’ test to see which of my own matches also match her. It is highly probable that the matches we have in common will be located on my father’s side. At least that halves the problem - they are not on my mother’s side. The converse is also the case - if I run the ‘not in common with’ test, the matches in the resulting list will probably lie on my mother’s side.

It’s actually a little more complicated with Katy, as she has one non-Jewish grandparent, and a Jewish father from a different geographical area to the ancestors we share. But I’m going to ignore those factors for the moment.

Last week I ran the ‘in common with’ test on CG. Katy does not appear, so CG is therefore on my mother’s side. And because he has only two identifiably Jewish great-great-grandparents - the Liskys - the connection must be through them. We haven’t got any further in identifying exactly where it lies, but at least we now know where to look.

FTDNA: Chromosome Browser
I next took a closer look at the ‘in common with’ results, in particular at the top 4 in the list. These are the people whose overall results place them closest to me, and who also match CG. Since they match CG and not my cousin Katy, they too will eventually turn out to be on my mother’s side. Again, I’ll refer to them by initials: MW, SLR, JW, DRT.

The Chromosome Browser test shows where you match each other on each chromosome, and how long the matching segments are. It also tells you how many segments you match on with each of the other people you have included in the test. 

Number of matching segments with me:
MW: 24
SLR: 23
JW: 20
CG: 11
DRT: 36

The point of interest here was that although CG only has one Jewish great-grandparent, as far as I can tell the others in this group all have 8, as do I. And yet despite having only 1/8 the amount of Ashkenazi genetic input, he nevertheless matches me on 1/2 the number of segments that most of the others do. I’m not sure what conclusions can be drawn from this. As we saw before, he also came up with roughly 1/2 the total shared cM as my other top 2nd-to-4ths - a different group of people (see previous section). Might it be that some bits of Ashkenazi DNA are more durable, or more sticky - less likely to mutate - than the average? Or maybe it’s just some of the bits that he and I share?

However, some of these segments are quite small, and may not be of much genetic significance.

I ran 2 checks on the Chromosome Browser, to see where we all matched. When I put the minimum match length at 5cM, there were only 3 points at which more than one person matched me exactly:

Chromosome 5: MW & CG
Chromosome 15: JW & DRT
Chromosome 18: SLR & DRT

The match on Chr 18 between myself, SLR & DRT is quite short, but the other two appear to be quite substantial. Chr 5 is where CG and I have our 26cM match, and MW substantially shares this. The same is the case on Chr 15, with myself, JW and DRT. These look like big chunks of DNA in common between 3 people.

I would like to conclude from this that, if and when CG and I manage to establish exactly where in our respective families our match is located, then we should expect to find one of MW’s ancestors there or thereabouts as well. Or of course the breakthrough may happen with MW’s family, pointing the way to the match with CG. The same thinking applies to the match on Chr 15 - JW and DRT are on my mother’s side, but maybe not in the same branch as CG and MW.

As I said, I would like to be able to come to this conclusion. Either that, or be shown why I shouldn’t.

I then set the bar lower, at 3cM. This gave me:

Chr 3: SLR & DRT
Chr 5: MW & CG
Chr 7: MW & JW
Chr 12: MW & SLR & CG
Chr 15: JW & DRT
Chr 18: SLR & DRT

The results in red are the new ones, and represent fairly short matches at between 3cM and 5cM. They are probably less significant than the longer matches. However, Chromosome 12 catches my eye - MW, SLR and CG all share a short segment there with me. Does that hint at a common ancestor for all 4 of us, somewhere way back in time? Notice that SLR seems to have joined the club formed on Chr 5 by MW, CG and myself, although she does not appear to have a significant match there herself. Could the Chr 12 group be related via a different common ancestor? CG only has 2 candidate ancestors - his Lisky g-grandparents. Could the Chr 5 match be via one of them, and the Chr 12 match via the other?

It feels like there are some tantalising hints here, if only we could work out what they are pointing us to.

GEDmatch to follow
CG has also submitted his DNA results to GEDmatch, as have I. The tests available there are different, and potentially very interesting. Next up. 

1 comment:

  1. Oy! This is all so confusing, but thanks for explaining the shared CM and the chromosome browser info/significance. Am in the process of slogging through all this information for myself. I'm staying tuned to see what you learn.