X DNA match on Gedmatch???

[LarryBurford]

I was contacted today by a woman in Australia who said my wife was one of her closest DNA matches on gedmatch. Here is the match.

Minimum threshold size to be included in total = 700 SNPs
Mismatch-bunching Limit = 350 SNPs
Minimum segment cM to be included in total = 7.0 cM


Chr Start Location End Location Centimorgans (cM) SNPs
X 119,288,160 146,925,995 45.2 3,402
Largest segment = 45.2 cM
Total of segments > 7 cM = 45.2 cM Actual.

What I do not understand is when I do the "one to one" or "one to many" matches this person does not come up in my wife's matches. The only way I get a match with her is when I to the "one to one X" and put both of their kit numbers in.

Can someone explain what is going on to me?

Larry

[ltd-jean-pull]

45cM is a good sized chunk of X-chromosome to match on.

I recently read this article by Roberta Estes

https://dna-explained.com/2017/01/19/concepts-segment-size-legitimate-and-false-matches/

It occurred to me that the section on false matches may explain why women have so many X-matches and men have so few. I would have thought that a woman would have twice as many matches as a man, since they have two chromosomes, but they have more like 10 or 20 fold the number of matches.

Have you tried reducing the cM length to a lower value - 3cM even. I'd also run a people who match both kits and see what happens.

I've recently seen that two relatives of mine who are third cousins to each other are not autosomal matches. I know that statistically this can happen, but it did make me question whether an extra generation had been hidden there somewhere (a grandchild raised as a child). But he is a stronger match to his cousins from his gt-grandmother's side than the rest of us, so I think not (whereas the non-matching 3rd cousin is related via the gt-grandfather).

[John McCoy]

The current way that X chromosome matches are detected by the "One-to-Many" tool on GEDmatch is apparently much too loose to be useful for females. In my experience, the VAST MAJORITY of such matches, especially for segment lengths below about 15 cM, cannot be confirmed by using the X version of the "One-to-One" tool on GEDmatch. In a few cases, I have been able to look at phased results for a female using the One-to-Many tool. In those cases, the number of X chromosome matches (the statistics at the bottom of the One-to-Many report for X chromosome matches are usually in the 10,000 to 20,000 range, in my experience), dropped to the 100-200 range, approximately, similar to what males typically get, and thosee matches appear to hold up using the One-to-One tool.

Therefore, I think the issue of finding vast numbers of false X chromosome matches for females on GEDmatch is simply an artifact of the algorithm that is used on that site in the One-to-Many tool. What that algorithm is, I don't know, but it needs to be improved.

[georgian1950]

Therefore, I think the issue of finding vast numbers of false X chromosome matches for females on GEDmatch is simply an artifact of the algorithm that is used on that site in the One-to-Many tool. What that algorithm is, I don't know, but it needs to be improved.

Except for the few at the top of the list that are related closer than about 4 generations, X-matches are not worth looking at. It's not the algorithm, but the interpretation. The results have a reason behind them, but it is something that most people cannot fathom.

Jack Wyatt

[gatty]

Except for the few at the top of the list that are related closer than about 4 generations, X-matches are not worth looking at. It's not the algorithm, but the interpretation. The results have a reason behind them, but it is something that most people cannot fathom.

Jack Wyatt

What is your reasoning behind this?

[mabrams]

What I do not understand is when I do the "one to one" or "one to many" matches this person does not come up in my wife's matches. The only way I get a match with her is when I to the "one to one X" and put both of their kit numbers in.

Can someone explain what is going on to me?

Larry

If you just do the regular "one to one" you should not expect her to show up, since that is autosomal. But you should see her on the "one to one X" and you do.

Not showing up on the "one to many" bothers me. I hope you dont mind me saying this, but just in case. You did sort the X column so that all the larger X matches float to the top, right?

[georgian1950]

What is your reasoning behind this?

OK, take your 'X one-to-many' list (if you are a male with only a handful of matches, you may have to borrow one from a female). Skip past the few close matches at the top and go down to where the ones with generations to MRCA predictions are showing around seven generations. Skip around and start running 'one-to-one' comparisons and save the data on the matching segments. Before long, you will start getting a lot of triangulations. Projecting out to the thousands and thousands of matches most people have you will find that you would have scores and scores of triangulations on specific segments. A woman cannot possibly produce enough offspring in seven or eight generations to achieve that level of common ancestry, so the common ancestor must be a male.

Actually these big matching segments are not IBD but are reconstructed segments of the common ancestor's DNA from having multiple lines back to him with each providing pieces of the DNA.

Basically nearly all of the matches on the typical person's 'X one-to-many' list are just showing that you share this common male ancestor around 300 years ago.

Jack Wyatt

[LarryBurford]

Thanks for all the replies.
Larry