Sunday, May 3, 2015

Some Oddish Results

I know that once you get to third and fourth cousins, autosomal matches get really iffy. According to the ISOGG wiki, third cousins share 0,781% of their DNA on average and fourth cousins share 0.195% on average. In terms of centiMorgans, that's 53.13 and 13.28 respectively. "On average" means of course that it can be more - or less.

Add to that mix the fact that Family Tree DNA's cut-off for acknowledging a match at all is nontrivial. You can have a small match but they don't count it. (I think their cut-off is 20 cM.)

Over the last couple of years, I have gotten used to Family Finder results that make some kind of sense and I have succeeded not badly at coming to conclusions that I did not think likely going in. This despite the vagaries of genetic inheritance that can show two fourth cousins who do not match at all, while their siblings match wonderfully and convincingly.

Another way of saying that is that I have gotten spoiled. That is why the newest compilation of Family Finder results feels so weird, though it really is not. For the most part.

Below is a table showing relationships among seven family members, identified by initials.

Here is the level of certainty of the relationships.
  • J, M and S are fully documented. 
  • F's relationship is based on a family tradition, confirmned by DNA testing (autosomal and Y).
  • G and R are fully documented. Their relationship with the others is based on strong naming patterns, supported by autosomal DNA.
  • D is related to the others based on naming patterns, supported by autosomal DNA.
The new results are for S and M.

The relationships on the top right are the suggested relationships according to the FTDNA matches. The bottom left are the actual relationships. Most of them are not bad.

D, the weakest of our assumptions, shows his first four suggested relationships as correct. F looks good, at least on the first three. G and R are great with each other and with J.

S's matches are generally is not as good as I'd like, though the matches with G, R and particularly D are important.

But the three fully documented relationships among J, M and S (marked in yellow) do not show up on FTDNA at all. This will undoubtedly raise credibility issues for the whole study, among some of the participants.

According to GEDmatch, the match between M and S is indeed very small, just 12.6 cM altogether with none of the three segments larger than 5 cM.

But the matches that J has with M and S on GEDmatch are not so simple.

Both are over 50 cM, so should certainly show up as matches on FTDNA. I asked the folks there to have a look at this and will add their findings at the end of this post when they become available.
The GEDmatch results are pretty much in line with the results predicted by ISOGG.

No comments:

Post a Comment