A Segment-ology TIDBIT
Bottom Line Up Front (BLUF):
1. Email GEDmatch Matches with Ancestry kits to identify their AncestryDNA Profile.
2. Use Ancestry tools to extend the Ancestry of DNA Matches at other companies.
As mentioned before, genetic genealogy is a combination of genealogy and DNA. Another way to look at it is: genealogy with a DNA tool. The foundation is our genealogy – we want to build it out and we want to get it right. That’s were the DNA comes in. Each segment of our DNA comes down to us from a specific line of our Ancestors to our mother or our father, and then to us. People who share the same DNA segment (called segment Triangulation), are related to us somewhere on the Ancestral line we got it from.
In the practice of genetic genealogy, we look for Common Ancestors (CAs) with our DNA Matches; and we look for DNA Matches who share the same DNA segment. Finding the same DNA segment can be done through DNA Painting, and/or by segment Triangulation (forming Triangulated Groups (TGs)), and/or, roughly, by Clustering. Because it is a precise process, segment Triangulation is generally considered to be the gold standard.
Our search is for DNA Matches with Common Ancestors AND specific segments.
The problem is that AncestryDNA does not reveal the needed DNA segment information. They have the largest database of DNA Matches, and, by far, the best Trees. Yes, many Matches have Private, or no, or skimpy Trees, but still: AncestryDNA has a much higher percentage of Matches with decent Trees than any other company. On the other hand, the other companies (23andMe, FamilyTreeDNA, MyHeritage and GEDmatch all provide the detailed shared DNA segment information, and tools to determine if these shared segments Triangulate. But the genealogy side of these companies pales in comparison to AncestryDNA – in numbers, quality and useful tools. For example I have identified over 4,400 CA-Matches at Ancestry, and only 575 CA-Matches at the other 3 companies (a few are Matches with multiple CAs). However, at each of the other 3 companies, I’ve grouped all my DNA Matches (with IBD (true) DNA segments) into my 372 Triangulated Groups. In other words, I have Common Ancestors for all 84 of my known 128 5xG grandparents (roughly 3/4 of my Tree at that generation); and 372 TGs that cover all of my DNA – if I could only link them together…
The effort now is to: 1) find segment data for Matches at AncestryDNA; and 2) find CAs for Matches in TGs at the other companies. The following two “tricks” have helped me a lot.
Trick 1. Email Matches at GEDmatch who have a DNA test from Ancestry. My standard email:
Hello – we share a DNA segment at GEDmatch (I am kit M200…), and I’d like to determine our Common Ancestor. I will do the research and report back to you on what I find. All I need from you is a link to your Profile at AncestryDNA – this could be your AncestryDNA user name and/or the URL of your Tree. My Ancestry user name is jimbartlett1; and my Tree URL is https://www.ancestry.com/family-tree/tree/20620230/
Hope to hear from you… Jim Bartlett email: jim4bartletts…
Please be sure to follow up and report back to each Match who cooperates.
Trick 2. Use Ancestry tools to extend the Ancestry of DNA Matches at other companies (where we already have segment data). Some of our Matches at these companies provide some information about their Ancestry – even a little bit may be enough. However, I cannot sugar coat this – it’s work! But Matches who share larger segments should be closer cousins – the Common Ancestors should not be very far back. For Matches in key TGs, I actually work on a quick and dirty list of their Ancestry – an Ahnentafel list – 2 parents, 4 grandparents, 8, 16… usually by this time I’ve picked up a probable thread, if not the actual CA. Sometimes the Match has true dead-ends and a CA cannot be found – but often a CA can be determined. More on this process in a recent blogpost here.
Neither of these two Tricks will guarantee success, but they may be helpful if you are Painting or Clustering or Triangulating or just researching your genealogy. I use them regularly. NB: To link a CA to a TG segment (mapping), we need corroborating evidence – certainly genealogy Triangulation with close cousins, and Walking the Ancestor Back for more distant ones.
[22BA] Two Tricks TIDBIT by Jim Bartlett 20210709
Hi Jim: Following your very instructive blogs for some time now. I have a triangulation conundrum that I would like your advice on. This is based on 23andme Relatives in Common data. I share a segment with MatchA. The both of us share the same segment with matchB and MatchC but the match lengths are different:
Self – MatchA : 10 cM
Self – MatchB : 14 cM
Self – MatchC: 10 cM
MatchA – MatchB: 10cM
MatchA – MatchC: 14cM
If we look at the overlaps, Self-MatchB extends ~2 cM in both directions from the 10cM overlap and the same happens with MatchA – MatchC.
I’m inclined to think that both 14cM lengths are IBD given the size. If so, how does one explain the obervations? Could it be an example of a false negative situation where Self-MatchA should match 14cM but some unique misreads/errors do not report it as such? If this is plausible, can misreads span 2cM? Thanks! River
The DNA segment you get from an Ancestor has fixed Start and End positions. Each of your shared Matches will also have a DNA segment from a Common Ancestor (that’s why you match), but each Match will almost certainly get a different DNA than you got – maybe starting sooner or later and ending sooner or later. What you “see” in a chromosome browser is the overlap between you and each Match. But the overlap will never include more than what you have (although the Matches compared to themselves may share somewhat more than you have). So the TG is formed by all the shared DNA segments – each one of which Matches most of the others – the TG starts with the earliest (smallest) shared segment start, and ends with the latest (largest) shared segment end. A TG is NOT confined to only the area where everyone Matches. Two additional points: 1. the data is fuzzy, so one TG may overlap the next one slightly (a few Mbp); 2. If a close cousin is in the TG the shared DNA segment may sometimes spread out over more that one TG – this takes some judgment to see where most of the Match-segments end.
In your example, you should sketch it out with actual start and end points…
Hope this helps, Jim
Thanks for the response Jim. My fundamental question is what makes a reported segment match smaller than it should be. Could no-calls, genotyping errors or other factors result in a reported segment that is 2 cM smaller than its actual length?
River, Yes. When I compare myself to my father at GEDmatch, there are several fairly small areas of no match (which is impossible). I’ve highlighted those areas in my spreadsheet, and in at least one case, a shared segment with a Match is split into two segments. Just like a diamond, the process does have some imperfections. Jim
Thank you for this very useful note. I was able to follow an earlier note describing a similar process using My Heritage data. I did not find it easy to follow the same process with GedMatch. Perhaps that is my lack of familiarity with that site. Can you suggest a few steps for downloading segment data on Ged Match? I am also using your suggestions for naming TG/s as they have evolved and extending Ancestry pedigrees.
The only way I know to “download” data at GEDmatch, is to run one of their reports (such as the Tier1 Triangulation; or the multikit analysis, or Cluster report). and then highlight the whole report, copy it, then paste it into a spreadsheet. That’s what I did long ago to start my spreadsheet. Now I just type the new kits in – one by one. Jim
I have enjoyed your notes and found them helpful. I did follow up on your earlier suggestions for using downloading and organizing My Heritage segment data bur have had some difficulty using GedMatch data in the same fashion. Have you outlined the process in a little more detail somewhere? I believe your suggestions for naming TGs has evolved. What is your latest recommendation? Thanks for your help in moving toward identifying TG’s. It is certainly a challenging Gold Standard.
Harold, My TG naming has stayed fairly standard: Example: 01S24 – means Chromosome 01; starting, roughly, at the S location (S is the 19th letter, so the TG starts between 180 and 190Mbp. Roughly – the idea is to keep the TGs in order in each Chr); 24 means I know its on my father’s , father’s  side – the 2 and the 4 are Ahnentafel number. You could just as easily use M (maternal) and P (paternal), in which case this TG would be 01SPP.
My work at GEDmatch is largely manual entry and comparisons – I already know the TGs and it’s just a task of deciding which side the TG is on and it falls into place.
Pingback: Best of the Genea-Blogs - Week of 4 to 10 July 2021 - Search My Tribe News
Jim,Enjoying some of your blogs. Wondering about your “tricks” here though. For the AncestryDNA matches lacking segment details, won’t you still be unable to accomplish triangulation (gold standard) even with a supportive response to your e-mails? Don’t you need to convince your matches to upload to another service to really succeed in placing and confirming the CA? I believe Ancestry is really really holding up accurate match validation. Thanks,James Carne