The mystery has been solved by JD Cowell, a member of the Visual Phasing group on Facebook! All GEDmatch kits were deleted shortly after on August 2nd, 2022.
This article was previously taken down and published again on June 8th, 2023 after a small paragraph at the end was removed and screenshots and names anonymized to respect the privacy of the people involved. With regard to the content, it’s still the last updated version of July 28th, 2022.
Number 17 happens to be my unlucky number, and once again it has proved to be just that.
When the DNA results of my father’s maternal half aunt V.P. arrived not that long ago, I was all over the moon. As a half aunt, she shares quite a bit of DNA with her nieces and nephews (on 20 out of 23 chromosomes!) and her results are incredibly valuable, because they point directly to their maternal grandfather. That amounts to gold, if you are a fan of the Visual Phasing technique.
I downloaded the newest version of Steven Fox’s magical excel spreadsheet, and set out to do Visual Phasing for my father and his siblings all over again. Thanks to this newly found half aunt, I was basically flying through all chromosomes until I reached that unlucky chromosome 17. It made absolutely no sense!
V.P.’s results helped me to distinguish the segments my father and his siblings inherited from their maternal grandparents, Ivan Permyakov (MGF) and Raisa Zenkova (MGM).
As for the paternal side of my father and his siblings, I’ve worked with the DNA of two first cousins from different lines – first cousin G.P. and first cousin A.S. Second cousins are much better for Visual Phasing than first cousins since first cousins have two grandparents in common, and thus, can’t help me to distinguish between my father’s paternal grandparents, Andrey Strelnikov (PGF) and Domna Pimshina (PGM). Nevertheless, they are still incredibly useful when it comes to finishing long chromosomes or assigning crossovers in those cases when siblings turn out to have very close recombination positions.
It isn’t the entire chromosome 17 that presents a problem, only the small region between 32Mbp and 54Mbp. My father’s sister L. shares a half-identical region (HIR) with my father and their sister G. there:
Sharing a half-identical region means that on one of her chromosome copies, either the maternal or the paternal, my father’s sister L. has inherited DNA from the same grandparent as my father and her sister G. And that doesn’t seem to be the case here.
On their maternal side, V.P. shares a 30cM segment with my father and G. from 32Mbp to 59Mbp and no DNA with L. This means that from 32Mbp to 59Mbp my father and his sister G. have inherited DNA from their maternal grandfather, Ivan Permyakov. Their sister L., on the other hand, must have inherited Zenkova DNA from their maternal grandmother at that region.
On the paternal side, L. shares a 60cM segment with her paternal first cousin G.P. from 9Mbp to 54Mbp. My father and his sister G. share no DNA with G.P. there.
I can’t say which paternal grandparent passed down his or her DNA to L. and her paternal first cousin G.P., but I can say that it wasn’t the same paternal grandparent, who passed down his or her DNA to my father and their sister G. at that region.
The same is true when the siblings are compared to their other paternal first cousin A.S.:
So, what is happening here?! If L. has received different DNA from different grandparents than her siblings on both chromosomes, then why is she reported to have a half-identical region with her siblings, instead of a non-identical one?
I’ve summarized the problem in the following image:
All GEDmatch kits were deleted on August 2nd, 2022.
And as if this wasn’t complicated enough already, let’s throw in a fourth sibling into the picture – their brother P.
At first glance, there doesn’t seem to be any trouble with crossover assignment, and it is possible to complete the chromosome. The comparison with V.P., however, throws everything off again!
Wait, there is more!
JD Cowell, a member of the Visual Phasing group on Facebook, has made a very valuable discovery! He checked the AYPR tool on GEDmatch and found that the results for my father’s brother P. were showing a green segment from 32Mbp to 48Mbp on chromosome 17! Now, isn’t that interesting?!
That could mean two things:
- my grandparents were very distantly related
- sibling P. has a partial trisomy 17 or a small uniparental segment
While the first explanation seems to be far more likely than the second, it doesn’t yet resolve the issue.
V.P. is the daughter of my great grandfather Ivan Permyakov and his second wife. Years ago, I’ve also tested her older half sister L.P., Ivan’s daughter from his first marriage to my great grandmother Raisa. As a full aunt, L.P. can match her nieces and nephews on both of her chromosome copies – which isn’t very useful for Visual Phasing (full aunts/uncles are only occasionally helpful with crossover assignment). But in a strange case like this one, let’s have a look at how the Strelnikov siblings match her, too.
In a normal situation, you would expect L.P. to match either the Zenkova or the Permyakov segments. However, two things are off in this image!
First, she doesn’t match my father’s sister L. from 39Mbp to 59 Mbp as she does with my father and sister G. – even though all three of them supposedly received their maternal grandmother’s (Zenkova’s) DNA segments in that region.
Second, their brother P. switches to a different maternal grandparent at 48Mbp, while my father and sister G. do not, but their full aunt L.P. is reported to share the same segment from 39Mbp to 59Mbp with all three of them.
Situation 1: Is this a case of endogamy or pedigree collapse?
My father’s newly half aunt V. is not related to any of the three of my father’s paternal first cousins, who have taken a DNA test, and their results have never contradicted themselves previously on any of the other chromosomes.
However, as I have mentioned earlier, group member JD Cowell has discovered that my father’s brother P. has one fully identical (homozygous) region on chromosome 17, which indicates that his parents were very distantly related. (As for the other three siblings, there was no such indication.)
Following JD Cowell’s argumentation, their maternal half aunt V. matches her half nieces and nephews on their paternal chromosome copy. As for my father’s brother P., P1 and Permyakov segments are identical from 32Mbp-48Mbp.
Following the same argumentation, the comparison between full aunt L.P. and the Strelnikov siblings is very similar:
Is sibling P.’s segment of 15cM enough to support this outcome?
Everybody comes across pedigree collapse at some point in their genealogy. And of course, it doesn’t contradict the history of initially only a small number of Europeans in Siberia, where my Russian ancestors settled centuries ago.
Then why do I still have trouble accepting it, even though it makes everything fall so beautifully into place?
There are several reasons:
Remember, the AYPR tool results for the three other siblings said “This analysis indicates that your parents are probably not related within recent generations”. P.’s segment is likely much much older than recent generations.
When I lower the thresholds for the comparison of my father’s newly found half aunt V.P. and all three of my father’s paternal cousins (G.P., A.S. and V.K.) who have taken a DNA test, the results are the following:
- 1 segment of 5cM shared with each G.P. and V.K.
- 2 segments of 4cM shared with A.S.
That’s not how relatedness looks like. Even if all of these small segments were real, the pedigree collapse would be hundreds of years old.
That contradicts the size of the segment V.P. shares with the Strelnikov siblings – it’s a solid one of 30cM. How are the odds of it being several hundred years old? According to a very interesting article by Kevin Borland, there is a 96% chance that a common ancestor of a 20cM segment lived within the last 400 years. And we are talking about 30cM.
When you lower the threshold at GEDmatch for the comparison of L.P. and my father’s paternal cousins, the number of segments is higher, but they are all still small except for one:
- 4 segments of less than 4cM shared with Galina
- 9 segments between 3 and 5cM and one segment of 23cM
- 11 segments between 3 and 5cM and one segment of 8cM
Again, even if some of them were real, the pedigree collapse would be hundreds of years old.
So, there is no recent relation between maternal and paternal sides. P.’s segment only points to a very distant pedigree collapse.
(Even if I compare L.P. to Blaine Bettinger, I still get 7 segments between 3 and 6cM 🙂 )
Situation 2: Could it be a case of hidden or double crossovers?
When I first visually phased the chromosomes of my father, his sister G. and his brother P., I tried different scenarios and could make the comparison work assuming several hidden close crossovers, as seen in this image:
However, everything is thrown off as soon as you take my father’s sister L. into the picture – nothing seems to explain the situation why L. is reported to share a half-identical region with her siblings when they are matching different people on both their maternal and paternal chromosome copies.
Situation 3: Could the 30cM-segment V.P. shares with my father and his sister G. be a false one?
This is extremely unlikely given that it is a rather solid segment shared with both siblings.
Situation 4: Does P. have a partial trisomy 17 or a small uniparental segment?
This is the unlikeliest theory of all. On very rare occasions, people can receive a small piece of a third chromosome or two copies from the same parent during the recombination process and the AYPR tool can give you a hint.
The argument against situation 4 is that there was an issue with this chromosome even before sibling P. was taken into the picture, remember? But P. has turned out to be crucial to the solution!
And that brings us back to JD Cowell. Has he solved the mystery of chromosome 17?
Maybe. Probably. Most likely. Yes. Definitely.
(Unless someone wants to challenge him? 😉 You can follow his arguments in the Visual Phasing group on Facebook.)