Thursday, February 20, 2014

Many More AncestryDNA Matches Now - and a Bug (Update: Fixed!)

AncestryDNA continues to provide more Hints to tree matches that also share significant DNA with me.  I last wrote about these matches in Happy to Have More AncestryDNA Hints  (15 October 2013).

I now have 3 matches who are 4th cousins (calculated by the trees, not by DNA) who are considered strong DNA matches (95% or higher) and 61 who are Distant cousins (5th to 8th cousins calculated by the trees), who are considered moderate to very low DNA matches.

Here is the top of my DNA match list with the "Hints" filter on(I wish it would tell me a number - I had to count to 64):

I won't show you the entire list to save bandwidth.

I have looked at all of the Hints on the list, except for the ones to Private Member Trees, and can see the connections to the common ancestors in my tree and the other person's tree.

The three 4th cousins are indeed 4th cousins, and I've corresponded with two of them off and on over the years.

I now have several matches that have more than one cousin link. Here's one example:

This match actually has three separate sets of common ancestors.  See the left and right arrows beside the common ancestor?  I can see the other sets of common ancestors, and the ancestral path from my tree and the other person's tree, by clicking on the arrows.

Match 1 of 3 on the screen above says that I am an 8th cousin to the other tested person.  However, the common ancestors are our 6th great-grandparents, so we should be noted as 7th cousins, not 8th cousins.

Here's Match 2 of 3 with this same tested person:

This match says that we are 8th cousins, although the common ancestors are my 7th great-grandparents, but there is one generation different, so we are really 7th cousins once removed.

Here is Match 3 of 3 with this same tested person: 

The common ancestors are the 7th great-grandparents of myself and the other tested person, and therefore we are 8th cousins through this match.

So what happened here?  Why does the AncestryDNA Hints tell me that I am an 8th cousin with the other tested person on all three matches with that person, when the tree lines say I am not?  The simple answer is that there's a bug in the AncestryDNA relationship calculator when there is more than one match for the two tested persons.  Apparently, the relationship calculator takes one of the (in this case) 3 common ancestors and calculates the relationship, and then uses that one relationship for the other matches with the tested person.  

I found this to happen for all cases where there is more than one set of common ancestors in the Hints for a tested person.  The relationships are calculated correctly when there is only one set of common ancestors.  In addition, my relationship with the other persons in the other tested person's tree are incorrect because of the error noted above.

I hope that will correct these errors as soon as possible.  

I REALLY like the presentation of these results.  By showing the relationships between me and the other tested person, I can now contact the other tested person using the Ancestry message system and compare noted on the ancestry of the common ancestors.  

The really neat thing is this:  if we find that we have significant chromosome segments that are identical, then that means our genealogy research back to the common ancestors is confirmed (unless one of us has left out or added a generation).  Unfortunately, doesn't provide a chromosome comparison between two different testers, but FamilyTreeDNA and 23andMe do. All three companies will provide a spreadsheet download of the markers and these can be read into the GEDMatch program to compare chromosome segments.  I have not explored that yet, but I will soon.

I hope that AncestryDNA will implement a chromosome matching comparison system soon. Ancestry has a pretty good thing going here - there are millions of Public Member Trees and a significant number of AncestryDNA autosomal DNA testers.  What's lacking is the charts that would show which part of specific chromosomes have identical segments. 

The URL for this post is:

Copyright (c) 2014, Randall J. Seaver

UPDATE:  25 February 2014;  I received an email from Ancestry that said the "bug" pointedo ut above had been fixed.  I looked at my example above, and the cousin relationships are now correct (not shown here).  This squeaky wheel is happy!


yuma mom said...

I agree with you about the chromosome browser. I have been contacting some of my matches and explaining the benefits of using Gedmatch. So far 4 people have used it and it has really helped me understand the match better.

Ginger Smith said...

Hi Randy, although I am not using AncestryDNA for DNA results, I have this problem with my RootsMagic and my Family Tree Maker Software. I haven't figured out any workarounds for it yet.

Joe Lowry said...

Randy, I noticed your shaky tree hint number is really high! Do you ever check your hints, even out of curiousity? How long have you gone without checking them to get the number that high? Just curious.

Unknown said...

Hi Randy,
I'm wondering if you've found a way to show DNA matches to a common ancestor in FTM? I'm using 2014 and am trying to find a way to show these relationships.