Autosomal DNA match anomaly or false positive?

Question

I've done autosomal DNA tests on ftdna for myself, my mother and paternal grandmother. I've transferred the results to MyHeritage, where I have a match with a person X with the largest segment of 27cM. Neither my mother nor grandmother match with X, so I assume it has to be a match from my paternal grandfather whom I have not tested. Looking at the match’s ancestory I can indeed identify common ancestors with my paternal grandfather. However, looking at the matching segment's location on the chromosome map, I see that it falls in an area where I also have a match with my paternal grandmother, which means that I cannot at the same time also have segments from my paternal grandfather in that same location. There's a clear contradiction there. Should I regard it as false positive or is some other explanation possible?

Edit: the following image is a screen capture from MyHeritage's chromosome browser. In red are my matching segments with X (6,3cM in cr 18, 7,9cM in cr 20 and 27,2cM in cr 22), in orange my matching segments with my paternal grandmother. As said, X and my paternal grandmother do not triangulate with each other. As seen, all of these three chromosomes (18, 20 and 22) are entirely from my paternal grandmother (the other pair of the same chromosome being from my mother), therefore there cannot also be any DNA from my paternal grandfather in the same chromosomes.

Update: The matching segment in chromosome 22 has a triangulation group consisting of possibly 10+ persons, including me, my paternal grandmother and X. Everyone triangulates with everyone, except my paternal grandmother and X, who both triangulate everyone else in the group, but not each other.

Answer 1

This question is a bit confusing. When you say "Neither my mother nor grandmother match with X", I'm assuming you mean your paternal grandmother.

You said "looking at the matching segment's location on the chromosome map, I see that it falls in an area where I also have a match with my paternal grandmother". You also said that your match with X is only 27 cM. The matches between you and X, and you and your grandmother, do not have to be exclusionary. In other words, your conclusion "means that I cannot at the same time also have segments from my paternal grandfather in that same location" is incorrect.

Possiblity 1) Your paternal grandparents are (perhaps distantly) related to each other, and share DNA in that position. You said that your grandmother does not match, but what we cannot see is whether there is zero match at all, or just a smaller match, between her and X at that position. A bad read anywhere within that segment in her test could have artificially prevented the segment from matching.

Possiblity 2) You're a bit unclear as to the "area" of the match. "it falls in an area where I also have a match" - is this talking about a) that entire chromosome, b) some part of it, or c) just the specific matching segment?

• a) the issue is comprehension; namely, you inherit bits of each chromosome from each parent. So even if your paternal grandparents are entirely unrelated, you could have bits from each of them within the same chromosome.
• b) similar to a, just on a smaller scale.
• c) that leaves only the possibility that you're saying that the entire specific segment matching between you and X is one in which you separately have a match with your paternal grandmother. However, that statement cannot be true without also saying that X and your grandmother have a match in the same position (unless possibility 1 above). If you and X have a match, then for that segment, you and X are interchangible. Logically, you cannot then have a match in the exact same spot with your paternal grandmother, and yet X not also have a match with her in that same position.

As for a false match, a 27 cM match is outside of any real possibility. In theory, sure, there might be some infinitesimal chance that you get exactly the wrong reads in enough places that a false positive of that size is created. That is so massively unlikely, however, as to be essentially impossible. A false positive would be expected to max out around 6 or 7 cM; a match of 27 cM is essentially certain to be a real match.

Let me just follow up on that last a little; just because it is a real match, it does not mean that it will be easy to identify the relationship between you and X. I have several people with matches in the 25 to 38 cM range that I've been working with for multiple years, where we each think we know roughly where the match between us should be in our respective trees, yet we've not yet found the connection. If this is the only sizeable matching segment between you and X, the connection could easily be 5C3R, 6C1R, 6C2R, 7C, 7C1R, 8C, and would require you both to have quite deep trees in exactly the right places to identify the common ancestor.

Answer 2

I'm unfamiliar with ftdna but consider that we are diplod organisms meaning you have two sets of each autosomal chromosome.