When a Plan Comes Together…

A Segment-ology TIDBIT

I love it. My prime objective with atDNA has been to map my genome to the most distant Most Recent Common Ancestors (MRCAs) that I can. The two essential ingredients for a Chromosome Map are Segments and Common Ancestors. So my basic game plan is to collect and Triangulate as many shared segments over 7cM as I can, and find as many MRCAs as I can. I have basically completed Triangulating the shared segments with all of my Matches (culling out many Identical By Chance (IBC) or false segments along the way), and now have 360 Triangulated Groups (TGs) covering 97% of my 45 chromosomes.

It’s now a full-court-press to find MRCAs with the Matches in these TGs. Of course, not all MRCAs will be correct, but the more I can find in each TG, the more data I have to develop and test possibilities.

Two ways to find MRCAs with segment data:

1. Start with MRCAs and get the Matches to test/upload to GEDmatch to determine shared segments [see Shared Ancestry Hints below], and

2. Start with Match segments and review their Trees (including getting them to share private Trees) [alas, so many Matches have no Tree.]

One process that has worked pretty well for me, focuses on AncestryDNA Hints. I have 830 Shared Ancestor Hints (SAHs), and I’ve sent a message to every one of them. It’s a standard message saying I agree with the Hint, but note that we might have other Common Ancestors, too. For that reason, and because I’m mapping my DNA segments to specific ancestral lines, I’d like for them to upload to GEDmatch so we can see the shared segment. It’s easy, and I will do the DNA analysis and give them a report back.  About 5-10% of these SAH Matches upload.

Today, in response to my request 7 months ago, I got a message with a GEDmatch kit# for a SAH who is a 5th cousins (5C). At GEDmatch I found our shared segment, typed the info into my Master spreadsheet, and Triangulated with other overlapping segments. The new segment was in one of my few remaining TGs with no known MRCA. So, from the Hint, I now had an MRCA! And it “fit” at the grandparent level with adjacent TGs. I then checked our Shared Matches – there were only 3 – one Private, one No Tree, and one with 57 people.  Well, the 57 people Tree had just the barest of a clue – a maiden name without dates or locations. But I knew where to look, and quickly determined it was the same line as my new GEDmatch kit. Wow! Identify an MRCA in a TG, and get another cousin with the same MRCA line in that TG at the same time.

I have over 600 Matches with MRCAs that “fit” at the grandparent level. And it’s becoming easier every day to find and rule potential MRCAs in or out of a TG.

Communicating with Matches to find MRCAs is the key. Sometimes it literally takes years… You’ll only get a small percentage of responses, but the more emails and messages you send out, the more you’ll get back.

Communicate!

[edited to identify abbreviations]

 

[22R] Segment-ology: When a Plan Comes Together TIDBIT; by Jim Bartlett 20180316

33 thoughts on “When a Plan Comes Together…

    • Sharon, I’ve taken the atDNA tests as soon as they’ve come out – 23andMe and FTDNA in 2010 and AncestryDNA in 2012. From 2010 to 2012, I focused on expanding my Tree to 13 generations wherever I could. And I communicated with every Match. And during those two years I determined about 100 Common Ancestors, while I learned about segments, and centiMorgans, etc. When I finally realized how to utilize the segment information, it became clear that about 1/3 of all my CAs were wrong! Well not wrong – we were still cousins on those CAs, but they could not be the CAs that passed down the DNA segment we shared. Even today, I still find about 1/3 of the MRCAs I find are on the wrong parent or grandparent. Anyway, as I developed my Triangulation process (collect segments, sort on Chr and Start, compare overlapping segments for Triangulation), I began to understand the big picture. By then I was fully retired, so I took about a week of solid work, and cranked through all of my shared segments – probably in 2014. I think it wound up with about 67% of a map. Since then I’ve been adding more and more from GEDmatch, and have gradually gotten to 97% of the segment map, with about 75% having MRCAs beyond a grandparent. I just added the MyHeritage segments to my spreadsheet and have over 10,000 rows. Starting from scratch, I’d recommend adding all segments over 7cM to a spreadsheet, but then sort out those over 15cM for Triangulation.

      Like

    • Doris – I had to retire, to even begin to handle all the data – and now I’m shifting to working mainly with 15cM segments and larger (I’ll drop down to 12cM if I can catch up on the 15cM category.

      Like

  1. Impressive use of abbreviations. Or code. I think I get it. It’s to keep people out of the club I guess. I don’t have to be told twice.

    Like

    • David – Sorry – I just edited the post. I’d like to be as clear as I can for genealogists – so I don’t want “code” or a “club”. There is a lot of concept to grasp, to understand segments and recombination/crossover points and Triangulation (see my other blog posts). I think segments and Triangulation provide valuable tools for the genealogist – but there is a steep learning curve. Jim

      Like

  2. Great job!

    I cant help thinking that people I ask to upload to gedmatch think I’m trying to sell them something even thought I try to emphasize that its free and fun.

    Like

      • Jim, do you have a template or hints for asking your matches to communicate & upload to GEDmatch? My success rate is minimal, I have tried all sorts of strategies, I get less than 10% reply, and less than that to upload, many feel that they are not computer literate enough to do it, or the rest think I want to clone them 😉

        Like

  3. Great post Jim!
    I have one other method that you didn’t mention: In 2015 I came across a 4th cousin match who was clearly a match on my brick walled Packard line. He and his sister and a cousin had all created very detailed trees out several generations. Only thing was (an all too common problem): Their trees contained only their direct ancestors, with generations of “One Child Families”.

    I asked the three cousins if they would mind if I created a private tree on Ancestry, giving each of them access, to build out their trees. The 4th cousin even gave me a GEDCOM file of his tree.
    That was how I discovered that their 2nd great-grandfather was Oakes Packard, and Oakes had a brother, Joshua Packard III, who turned out to be MY 2nd great-grandfather. Joshua was the great-grandfather of my brick wall ancestor and the wall came tumbling down.

    I’ve been asking responding matches if they mind if I create a private tree ever since.
    Granted, it’s nice if they have a tree, and it’s detailed and all built out. But if they don’t, a good and dedicated genetic genealogist can create one with some information provided by a cooperating match. You won’t know about the opportunity if you don’t ask.

    Like

  4. Hi Jim,
    I really envy you your ability to make a plan and go for it. I have followed your writing for a long time and really look forward to reading them, and over again.
    Kimmer

    Like

  5. 830 shared ancestor hints? I have 59. 😢 I only wish I could build out most of my tree to 10-13 generations. Most of my lines seem never to have been researched before….

    Like

    • I can relate. I have three adopted daughters, and I have half to a quarter of their matches. Either Italians are slow to test, or being in the country only a decade over 100 years has meant limited family size. But then I have no excuse to not have contacted EVERY one, right?

      Like

      • The sad thing is that my response rate for ancestry messages is very low. FTDNA is better, where we have actual email addresses. I think many people don’t even know they get messages via ancestry. I do know where some of the ancestry matches fit into my tree, yet not being able to see the segments really hampers me in making further progress. If they don’t see or won’t respond to my messages, I have no hope of them uploading to gedmatch. Ancestry is really a struggle. With FTDNA, even if they won’t respond, I can at least see the segment data.

        Liked by 1 person

    • Valerie, I use Genome Mate Pro. High learning curve on the front end, but helpful for organizing chromosome matches from varied testing companies. There is a way to back end Ancestry matches, if you have a PC. I have not been able to rig it for Ancestry to download matches and their data, but if you are interested I can send you links.

      Like

      • That would be awesome! Thank you! I have a whole group of matches on one of my mom’s lines (with major brick walls). There are one or two people they all match, but I’ve only gotten a response from one in the group so far. 😦

        Like

      • Thanks! I’ll take a look. I’ve heard of it before but haven’t used it. I generally just use excel spreadsheets for this, but obviously that doesn’t work with ancestry.

        Like

      • Valerie, GMP and DNAgedcom Client and several 3rd parties provide different tools. The ISOGG Wiki, https://isogg.org/wiki/Autosomal_DNA_tools provides a long list of tools. The Match info and the segment info and the other things you might want to track and manipulate are all just data – I like a wide spreadsheet, with columns for any data I want to track (and am willing to input…). Mainly I sort by Chromosome and Start, which puts overlapping segments close to each other and convenient for Triangulation. Once the TG are formed and assigned to a parent, I can then sort first by the parent and then by Chromosome and Start, and all the TGs and Match-segments in them, are displayed for each chromosome. I then work within any TG I want, with whichever Matches I want, to find MRCAs – there is a lot of synergy in this. Some of the other tools do the same things. The main takeaway is that the segments are like jigsaw puzzle pieces – they only fit in the big picture one way – no matter which tool(s) you decide to use. I can visualize the segment from the segment data, others prefer tools like DNAPainter, or Kitty’s Chromosome Mapper, our DoubleMatch Triangulation, or … to help visualize the how stuff fits into the big picture. Whatever works best for each person – there is no “ultimate” tool (although I think spreadsheets come close). Actually what we should be using is a relational database, but, for me, the learning curve is too great to start over… Jim

        Like

      • Jim,

        Yes, I currently use an excel spreadsheet, started with my data from FTDNA, and I load in individual matches from gedmatch and myHeritage as I’m able. Ancestry is just a problem. The biggest database–but no segment data. I can tell a lot in some cases by who the common matches are, but until several of the people in a group upload to gedmatch or somewhere to give me segment data, it’s impossible to tell for sure who all overlaps (and matches) who else. So I don’t get real triangulation until that happens.

        Ancestry is great for seeing images of documents, and they have a huge database–but they give us a lot of challenges in analyzing DNA.

        Valerie

        Like

  6. Jim
    Having come to genealogy, and then your posts much too late, I often find myself at sea without a paddle? Clearly tools are critical, and I was going to ask if you could publish your worksheets.
    I am very glad to read that you will publish your work on spreadsheets!
    A suggestion, can you include the initial underlying building blocks for your spreadsheets ?

    Like

    • Don, Yes. The three fundamental blocks are: 1) information about the Match (name, email, Tree, etc.); 2) information about shared segments (Chr, Start, End, cMs, SNPs); and 3) other information that you add as you go along (CA/MRCA, exact Cousinship, close kin who also match on the same segment, Triangulated Group ID, Send and Receive info, Remarks, etc.) The first two block are generally included in downloads from your testing company(s). Jim

      Like

  7. Jim, I am rather new to segment analysis and am using Genome Mate Pro and GEDmatch. I have a mystery that I wonder if you have written about. I have three shared matches of over 75 CM’s with one large segment of 70 CM’s. The SNP’s are close to 9000. I have never read anything about shared segment size, but imagine that a segment this large should mean something. I am using this tool to pinpoint possible relationships (https://dnapainter.com/tools/sharedcm), but I am still sitting with this strange situation: I have not had a DNA match with one segment so large, that was not a clear, known family member. Can you help me figure this out (I have some theories which I am also happy to share).

    dm

    Like

    • Dawn – the larger the shared segments, and the Triangulated Groups, the closer the Common Ancestor will be – in general. TGs that are 70cM may well break up into smaller TGs with more distant CAs. I think one of the best strategies in this case is to try to set up a conversation with multiple Matches, to encourage them them to compare among themselves. They may well come up with something you don’t have. Jim

      Like

      • Thank you, Jim. We are in touch and I have done extensive research for multiple trees. The 70 CM is the largest segment, which is why I posted the question. It is one of my matches largest segments, yet we cannot find the MRCA. I think this is leading to a patronage that is currently unknown or undisclosed to our families.

        Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s