Segments with no matches

Question

I am beginning the process of forming triangulation groups, and I wonder: what do people do with segments on which there are no matches?

I found such a segment already just on the first chromosome I have examined. There is a region with no matches over the threshold on any of FTDNA, MyHeritage, or GEDmatch.

I can think of a couple of reasons: * going back many, many generations and following this segment through my ancestors, it just so happened by chance that all other descendants had this segment excised at some point through recombination (or at least, all descendants that tested DNA at these places) * a mutation happened recently, close enough to another mutation to cause otherwise matching segments to be flagged as non-matches.

Is one of these the more likely explanation?

Should I view short matching segments in such a region as more reliable, i.e., more likely to be IBD, say because it is a personal "anti-pileup region"?

Answer 1

There are a few different reasons why you may have some segments that have no triangulations or no matches.

1. It could be, as you state, that the segment died out over time for all your relatives and you were the only one who got it. This could be because of the random 50:50 chance of it passing on at each generation and after 4 or 5 generations, your relatives got the "unlucky" rolls.
2. More likely is that the segment was passed down to some of your relatives but maybe something like only 1 in 20 of your relatives got it, and none of them happened to have DNA tested.
3. Also possible is that your ancestor's line mostly died out, with the extreme case being that each generation had only 1 child who then went on to have a child, ultimately leading to you.
4. It is very unlikely to be due to a mutation, because the autosomal mutation rate is so low that there are only on average 175 mutations a generation. That's out of 3 billion base pairs. So in 700,000 or so tested SNPs that are tested by the DNA companies the expected number of mutations per generation would be 0.04. More likely than mutations, but what would cause the same problem as you hypothesize in your question, are read errors in the DNA test. I have not seen published read error rates, but my own comparison of FTDNA and MyHeritage DNA tests found 42 differences among the 700,000 SNPs. On average then, they would be 85 centimorgans (cM) apart from each other, so there is a small possibility that two will be close enough together to not allow the segment to match, but it is unlikely to be a major cause of the gaps.
5. What I think is the most likely reason for lack of matches or gaps between triangulation groups is the particular place that the recombinations occurred. If, say, in your DNA, you have a crossover of your parents at a particular place on a chromosome, and your father had a crossover of his parents within a few cM of that, and your mother also had one from her parents within a few cM of that, then you will happen to have a mix of all four of your grandparents over a short segment, and you won't have any relatives who will be able to match that. It can be with grandparents as well, which will prevent 1st cousins from matching you. Or with great-grandparents, preventing 2nd cousins from matching you. I would expect most matchless and non-triangulating regions are due to the close crossovers of you and/or your more recent ancestors that passed their DNA to you.

If you have non-triangulating regions where you cannot build triangulation groups, but you do have short matching segments in the region, then remember that like all segment matches, those segments can only be one of three possibilities:

1. IBD (Identical by Descent) on the paternal chromosome.
2. IBD on the maternal chromosome.
3. A false match that is not IBD.

If you have more than two segments and none of them match each other, then at most only 2 can be IBD, one on the paternal and one on the maternal chromosome, and the rest have to be false.

With regards to whether a short matching segment in a non-triangulating area is reliable, treat it like any single match that does not triangulate. If it is above 15 cM, it is likely IBD, but if it is below 15 cM, it may be false and you'll need to find more evidence or other matching segments to indicate it is IBD.