About

Jim Bartlett has been a genealogist since 1974. He started the BARTLETT DNA Project at FamilyTreeDNA in 2002. Soon thereafter he began teaching DNA to various genealogy groups in MD, VA, DC, WV, etc.  In 2010 Jim took his first autosomal DNA test at FamilyTreeDNA, followed shortly by atDNA tests at 23andMe and AncestryDNA. Jim has become a huge fan of the Triangulation process and chromosome mapping – the ultimate human puzzle. This site was set up to document information about segments and autosomal DNA.

47 thoughts on “About”

  1. Pingback: Genealogy News and Reflections of 2021 – The Genealogy Situation Room

    • Jim
      I have followed you for quite some time and have used you Segmentology processes. I have been quite busy in the last few months with the 7 Ancestry DNA profiles I manage and haven’t had time to work on my 23andme ones. My Xcell spread sheet have gotten almost to large to manage. Has anyone developed a tool/data base that would allow the management of the vast detail we need .
      Doug

      Like

      Reply ↓

  4. Jim, I saw a recent post (on some Facebook group, I think) where your “Roadmap” was posted. Unfortunately, I copied the wrong link and now can’t find it again. It is an excellent step by step guide, especially for people new to testing. Can you please direct me to it?

    Thank you for all you do!

    Doris

    Like

    Reply ↓

  5. Jim – I often recommend your blog to other people trying to get started in genetic genealogy. I just noticed a couple of links are broken in your Outline of Segmentology – specifically 03A (Fuzzy Data) and 21A (Why upload to GEDmatch or FTDNA). The posts are still there if I search for them in the handy search box that comes up with the “this is embarrassing” response to clicking on the links. Thanks!

    Like

    Reply ↓

  6. Hi Jim my name is Michael and I got a 1st cousin match on ancestry of 821cms over 30segments I’m 93% Ashkenazi Jewish and this match is 83years old I’m in my 30’s. Is this match a true 1st. cousin? your help would be greatly. appreciated.

    regards
    Michael

    Like

    Reply ↓

  7. I have a segment question for you. I have a match to a 3rd cousin 1x removed on my paternal side. My father, my 3 paternal siblings, myself and my son all share segments with this gentleman on several Chromosome segments. Obviously due to recombination, we do not all share the same number and ranges of chromosome segments which I understand. But my question is, why would the start/stop range be larger for a child to this cousin than their parent to the cousin? Especially when there is no reason to believe that the cousin matches both a paternal and maternal relative due to IBD or IBS. Is it due to a mutation? The match is from an Austrian match, my husband and son’s father is 100% Chinese.

    Like

    Reply ↓

  8. Jim
    Based on your methodology and what I have read on DNAGEDCOM.com I jumped into the deep end of the segment-ology pool!
    I manage my family’s history and have a large tree on Ancestry.com (10,000 people with over 1,000 DNA matches for the DNA kits I manage) and have had several of my family take DNA tests (11 on Ancestry, 4 on 23andme.com, 1 on FamilytreeDNA). I have loaded all but 3 on GEDMATCH.COM and am in the process of utilizing DNAGedcom.com as a way to normalize the data.
    I have utilized DNAGedcom for the ADSA analysis to obtain In Common With information (ICW) on my Montenegrin cousin’s GEDMATCH data (Dimitrije Poleksic). I created an ADSA output Excel spreadsheet and modified it to conform to your recommendations for a triangulation group sheet.
    I now have gone thru the spreadsheet and created triangulation groups following to your steps. (the idea of dividing the start and end segment values by dividing by 100,000 made it a lot easier)
    I picked Dimitrie as a 1st step as he has the smallest set of data and before I tackle my data I want to make sure I have the method understood correctly.
    Is there a way I can have you take a quick look at my work to date before I proceed?
    Thanks again for sharing.
    Doug Saucier
    VDS5045@gmail.com
    P.S.
    I am also a retired IT guy.

    Like

    Reply ↓

    • Doug, I’ve sent you an email. In general, I don’t want to take the time out to analyze everyone’s spreadsheet. I cannot keep up with my own. But if you create a spreadsheet and form TGs, you’ll find that new Matches will match segments in one TG and not the other – this is the ultimate test! IBD segments have to go on one side or the other. Jim

      Like

      Reply ↓

  9. Jim
    After 2 years of working with aDNA date and struggling with the concepts of matching segments, triangulation, and making spread sheets I found your site and want to thanks you for your comprehensive set of information. This site has cleared up a lot for me.
    Thanks again
    Doug Saucier

    Like

    Reply ↓

  10. Hi Jim,

    I’m so glad I found your blog! I saw a reference to this page with the GEDmatch triangulation tool. When I started doing DNA testing, it wasn’t very long before I started asking why everything wasn’t geared towards mapping your chromosomes. I’m somewhat frustrated with the tools that are currently available for DNA analysis. It would be so great if Ancestry would take the lead and not only give us a chromosome browser (PLEASE!), but also facilitate building a chromosome map and allow that to be shared with others. I have mixed feelings about Ancestry, but they have a large database and the tools to build a document-based tree. It would be amazing if they would have some kind of wizard in place for helping casual users get more out of their data and help the community at large, instead of their DNA circles that have essentially no transparency. As we find new matches it would be so helpful if one (or both) of us were in a position to say that this particular segment goes with this ancestor in my tree, instead of combing through branch after branch of stranger’s trees (when we are lucky enough to find an attached tree!). I imagine a system where the DNA and family trees are fully integrated such that you click on an ancestor in your tree and you see all mapped DNA segments (from you and others) and you click on a DNA segment and it takes you to your family tree and shows all DNA matches with their lines. Of course, the issue is the balance between privacy and the kinds of tools that would make this work faster and easier.

    I have a question for you: can you tell me why tools like GEDmatch’s matching segment search or FTDNA’s chromosome browser don’t provide phase information? It would save me a lot of time if I could see right away whether segments at the same chromosome position were the same phase (i.e. matching). As you describe in your blog, there are tools for seeing if the matching segments from two people match each other and are therefore assumed to come from the same ancestor. But why do we need to do the second step? Why can’t we see some kind of alignment analysis right there and then?

    Thanks!
    Rich Dixon

    Like

    Reply ↓

    • Rich – Thanks for your comments – I agree it would be nice for ANY of the companies to provide Triangulation tools – particularly ones which could link a TG to a Tree.
      GEDmatch does provide phasing tools.

      Like

      Reply ↓

  11. Hello Jim, I saw a link to this site after expanding my activity of FB, and have been inspired to take action with my autosomal DNA. I have had some success with using Y DNA and have been able to successfully trace and then prove my research after a speculative period based only on best guess, as all I have to go on, on my paternal side was a registry office wedding of my gt grand father in Portsmouth, England, naming his father. I researched all the father son combinations in the Irish records of the relative, guessed, DOB, and eventually settled on one Barry family. Along the way a death certificate was discovered which names Patrick as the person dealing with his mother’s burial and registration in 1879. I eventually found a marriage certificate in Dublin for Patrick which named the parents. Inspired by this find I managed to trace all his descendants to the present day in Ireland and England, but there was still the BIG question, have I found the right family? Luckily Liam M Barry a young man living in Bolton, England was enthusiastic about helping after we exchanged Emails for a while, we agreed to meet and he would supply a sample. As I have extensive Y DNA test inc. Big Y at FTDNA it was exiting to find that our 37 STR markers were a close match 36/37, so I proceeded with selective SNP tests which all matched me exactly, So this proved him to be my third cousin and Garrett Barry DOB abt. 1835 to be our common ancestor. I followed up with autosomal testing and we have a reasonable level of matching segments. I have sort matches at FTDNA, 23andme and from Gedmatch, and have had several tries to tabulate matches, but there was something missing that I didn’t get the results I was wanting, as I still don’t know DOB and location of birth, OR marriage details of Garrett & Catherine Toner, and this represents my brick wall at the moment. I have every confidence that DNA testing of one form or another will get me to the next link in the story, so when I saw your methods I was inspired to try again, as I have used Excel extensively for recording BMD and extracting census information. The main problem I have is that my father died in 1969 whilst I was a teenager, and my mother who started our family research was yet to make a beginning. Luckily I have tested my mother and sister, and my Uncle , Mum’s brother, and through Gedmatch I have a phased extracted signature for my father. Using this I have searched one to many, and then individually extracted the segment matches as per your method, sorted by CH and segment starts, and have the start of a research project. I have only used the Gedmatch data so far and intend to do a similar exercise at FTDNA and 23andme, and when my next results arrive at Ancestry I hope to collate their matches too, although I note that they do not have tools available as do others. I figure the more exposure the better the chances of finding someone who will help clear the fog!
    One thing I have noticed is that every site uses different levels of cM to produce their matching list, and this can give deceptive levels of apparent matching. I had an experience with myheritage recently where I appeared to be getting abnormally high levels of matching cM, AND over multiple CH, but after comparing myself and my third cousin found just how low there threshold is and hence the high amount of apparent matching and cM.
    I’m not 100% sure about the next steps with the excel document, so will write again with what I need to do next and how to do it, that is making up the TG groups and how to research the nest stage?
    Thank you
    Mike Barry

    Like

    Reply ↓

    • Mike

      I have found that Triangulating every segment over 7cM creates a pretty comprehensive chromosome map. And encouraging as many of the closer Matches at AncestryDNA to upload to Gedmatch; and asking GEDmatch’s A kits their Ancestry name (or link to a Tree) is very helpful. Then work with Matches with shared segments over 10cM in each TG to search for the COmmon Ancestry. After a chromosome map; it’s mostly genealogy work.

      Like

      Reply ↓

  12. I love your Triangualtion group tool on gedmatch but have two question on reading the output. Each box has a kit number and name and underneath something like C1 770.6cM which I assume is the chromose with overlap plus the length of the overlap in cM. Below that is something like O(0,0) > 1(0,1) which I don’t understand.

    Second question is these boxes may be in a horizonal line or dropped down into a tree and I don’t understand what a drop down means or the gaphic at the extreme left. When in a straight line I assume that it is a set of linear descendents from left to right.

    Your help on these greast additions would be greatly appreciated.

    mike

    Like

    Reply ↓

    • Mike

      Disregard the (0,0) numbers – they are programming code as to where the box goes. I think horizontal Matches share most of a TG segment; drop downs share part, and don’t necessarily match everyone in the TG.

      Like

      Reply ↓

  13. Hi Jim Mike here again here is some background information on the same relative I asked about previously but here is more complete data from FTDNA Shared Segments: 28, Chromo 22 38093010 45772802 16.3 2291, Chromo 15 24541779 32194759 10.74 1880, Chromo 15 38407035 72794853 37.19 8550, Chromo 8 6698105 26648327 30.45 7228. Your help is greatly appreciated.

    Regards
    Mike

    Like

    Reply ↓

  14. Hi Jim my name is Mike and I’m of Ashkenazi Jewish decent my best match is on Gedmatch/FTDNA at 2nd cousin is this match an actual 2nd cousin? .Here are his stats on Gedmatch 154.5 36.8 3.3 chromo 8 6674201 – 26748284 (32.7244 cM) , Chromo 15 24321650 – 32548787 (20.1942 cM), 38407035 – 72820453 (36.817 cM) , Chromo 22 38088487 – 49528625 (27.5137 cM) 38088487 – 49528625 (27.5137 cM). Your assistance would be greatly appreciated.

    Regards
    Mike

    Like

    Reply ↓

  15. Jim:

    My husband recently found out he was adopted (at age 55). He has tested at Ancestry, 23andme, FTDNA (both mtDNA and YDNA 67 test). We have uploaded his 23andme and Ancestry results to GEDmatch. On Ancestry he has two 3rd cousin matches who are not related to each other. On GEDmatch his closest match is at 3.9 generations. Since we are neophytes with all this ancestry stuff and do not have a family tree to compare to, how would you suggest we locate a common ancestor? Any other suggestions you may have would also be greatly appreciated.

    Thanks,
    Valerie

    Like

    Reply ↓

  16. Although an ICW utility starts with everyone who is already a match (somewhere), the ICW grouping is just built on genealogy – no other DNA input. So an ICW grouping shows groups with the same ancestry, sometimes the link is on different segments. I’ve found this even between sibling Matches. And other times, notwithstanding the ICW group, the DNA linkage turns out to be on a different ancestral line. The ICW groups form good starting points, but there is too much random variation (for me, anyway) to assume they share the same segment. Jim

    Like

    Reply ↓

  17. Jim,
    I just found your site referenced on Gedmatch. It looks great.
    I have been working on a way to identify shared segment groups in Ancestry matches. It appears that it should work.
    The basic idea is that we can see what must be shared segment groups by just looking at the “In Common With” data. For example, I might see that Bob, Jane, Bill, and Mary all match each other as well as me. The simplest explanation is that we all have the same shared segment.
    So I can pull out all of my shared match data using the DNAGedcom tool and mine the data for groups that overlap with each other. There are fancy cluster analysis tools for doing this but I have found an Excel pivot table is a good start.
    There are three problems, though.
    1) If I match someone in more than one segment, they will be members of more than one shared cluster group and it will appear that the groups are actually a single group. This is easy to solve- I start by removing all of the individuals who match me in more than one segment.
    2) I match Bob at one segment and I match Mary at another, but when they match each other at a third the two groups will appear to be joined. Most of the time this will be because we are all descended from a CA who passed on all three segments to their descendants in different combinations. This is not always a bad thing, since if I am sure one group descends from a particular CA then the other is linked into that line also.
    3) I match Bob at one segment and I match Mary at another, but they match each other at a third due to a CA who is not related to me. Imagine if one of my paternal cousins married a maternal cousin. Their child would match people on both sides of my tree who really had no ancestors in common.
    The last two problems are more difficult to deal with, but not impossible. In each case there will be two subgroups whose members are highly linked, and then there will be one or two individuals in each subgroup who link to each other and combine the groups. In practice it will be relatively easy to identify these cases. One special case to watch out for is when one administrator has kits in two subgroups- in that case it is almost certain that the link is due to their being close relatives.
    A final point is that Ancestry only reports “In Common With” for 4th cousin or closer. So Bob, Mary, and I may share a segment, but if Bob is a Distant Cousin he will not show up in Mary’s ICW matches, and if they are both Distant Cousins neither will show up in the ICW matches. For now I am including these Distant Cousins and just need to realize that the absence of a ICW match between them is not significant.
    I applied this to my mother’s 15,712 matches. I excluded all of the dyads who just match each other and ended up with 1160 people who match me in one segment and have multiple ICWs. My first attempt gave me about thirty groups. The biggest group has 160 members and has many known relatives descended from Phillip Davis and Margaret Reed. About half of the other groups were also easy to assign to known lines. I haven’t gone to Gedmatch yet to figure out what segment each group corresponds to, but that should be productive.

    Like

    Reply ↓

    • Thanks, Andrew – I blogged on a method to use AncestryDNA, but it included encouraging them to upload to GEDmatch. I found over 50 Matches who all had my 7G grandfather in their Tree. And some of them were in Shared Matches together. But the segment data at GEDmatch showed about 6 of them in one group (on one segment, in one Triangulated Group); 4 of them in another TG; 3 in another TG; and the rest scattered as “onesies”. So your method is a good start – but the real grouping can only be done, IMO, with segment data. NB: I believe AncestryDNA will show a Shared Matches list for every Match; it’s just that the Shared Matches themselves will only be 4C or closer (which really actually includes a lot of true 5C and 6C in their 4C category). Anyway, you can “back into” many more Shared Matches by checking the more distant Matches one-by-one – it’s actually helpful to know this “back door” info and put it in the notes. Jim

      Like

      Reply ↓

      • I read your Ancestry post which is why I posted the above.
        Being aware that Ancestry ICWs cluster into shared segment groups means that you don’t have to contact as many people. If you get a couple of them to post their file on Gedmatch and the segments match you can be reasonably confident that the rest of the group shares the same segment.
        I have several separate groups that each shares a different Davis segment. So, yes, just because a bunch of people are descended from a distant ancestor and are my DNA matches does not mean that they all share the same segment(s) with me or each other.
        I didn’t mention it, but you can also go back and look to see if the people you share multiple segments with are members of the shared segment groups. Since these are closer relatives and presumably easier to find CAs with it can help assign groups to specific lineages.
        Ancestry shows ICWs of 4th cousin or closer for every match. It never shows ICWs of Distant Cousins for any match. So it is fairly common to find a one-way ICW. (Bob is a 4th cousin, Jane is a Distant Cousin. Bob shows up as ICW when viewing Jane’s profile but Jane does not show up when viewing Bob’s.) The DNAGedcom tool is the way to go- a few hours of it cranking away and you have the entire set of ICWs. I don’t have any connection with them- I’m just a satisfied user.
        Anyway, this is yet another complementary strategy. One can never have too many tools in one’s kit.

        Like

    • Rebecca, Thanks for the link to your work. Just about everything you say in your post is the same thing I’ve been saying for years – and so I like your article and the direction you are emphasizing. Many good pieces of advice in your article.
      If you’ve read all of my blog posts, you’ll understand that I think Triangulation and Triangulated Groups (TGs) are just fine. In fact one could substitute TG for IBD in your Article. My TGs are the same as your IBDs, with all of the same rules as well as cautions (many of which I’ve similarly stated). There are two types of Triangulation: genetic (or segment) Triangulation (TGs formed from sufficiently overlapping shared segments from widely space cousins – this is a purely mechanical process and does not require genealogy); and genealogy (or ancestor) Triangulation (getting at least 3 widely spaced cousins to agree on the Common Ancestor (CA) – and in the end getting most of the Matches in a TG to agree on the same CA line). Since TGs have to be on either a maternal or paternal chromosome, some genealogy is required to assign them to the proper side. Triangulation does not lead to errors, it leads to rock solid chromosome mapping. Any process (including “solving IBDs”), if done incorrectly can lead to errors. I’ve been a very active genealogist since 1974 – and an active genetic genealogist since 2002. My atDNA spreadsheet now has 259 TGs with CAs, 102 TGs w/o a CA, and 55 “fabricated” TGs in the gaps to “fill up” all 45 of my chromosomes. I have many TGs with multiple CAs identified with the Matches; but only one ancestral line is linked to each TG – that is one of my ancestors, at each generation, going back to the “founder” of the TG. There is, in fact, a single ancestor who started each full TG. Ancestral to that, the TG does not exist intact, it was different, smaller segments from different ancestral lines. Some of our Matches, with smaller shared segments can, however, descend from more distant ancestors of that CA (but not with the full TG segment) – this leads to a subset TG with it’s own, more distant, CA. Our TG Matches can also have MRCAs with us all along our own ancestral path. This is why “walking the ancestor back” is an important method to “solving” a TG (along with genealogy Triangulation). With all of the new Matches that are now (2017) pouring in, we should be getting cousins at all levels in our TGs.
      Thanks again for your comprehensive article that applies equally to solving IBDs/TGs.

      Like

      Reply ↓

  19. Do you analyze GEDmatch results for a fee? If yes, what is the hourly rate? If no, can you recommend someone? I am trying to analyze my results to identify my great great grandfather. Y-DNA testing has not helped us. Three of us on GEDmatch are 2nd cousins (we are all women) and there are more cousins from different children of my great grandfather on GEDmatch. With three 2nd cousins, in theory shouldn’t it be easy to zero in on the common ancestor who is our great great grandfather? The problem is the analysis as I am finding it difficult to do.

    Like

    Reply ↓

    • Maralaina,
      I am fully retired and devote most of my spare time to DNA and genealogy. One of my goals is to teach others how to analyze atDNA data. I have spent the last 6 years analyzing mine and am now working on my father’s atDNA. To do it all for a person takes a major time investment. I don’t charge for what I do, but I also don’t analyze other’s DNA. You might try http://www.ISOGG.org/wiki – click on the autosomal TAB and look for the list of resources. However, you are on the right track. Find all the shared segments (from your Match list), which also match (Triangulate) with your cousins from your great grandfather. In theory half of those Triangulated Groups will be on your great grandfather’s maternal side, half of them will be on his paternal side. Work with all of those Matches to see where they have a Common Ancestor. At this point it’s all about the genealogy – the DNA can only help so far.
      Jim

      Like

      Reply ↓

  21. Jim,

    I am having trouble understanding a particular nomenclature in the TG Tree that GEDmatch built for me. I have searched through your explanations in this blog and have been unable to find an answer.

    In the boxes with the red outline I understand everything but the bottom line. An example is “3(0,1)>4(1,2).” Can you help me with this?

    Thanks,

    Jim Campbell

    Like

    Reply ↓

    • Jim – please disregard those cryptic numbers in the bottom row. They were used by the programmer to help set up the graphic. They were used to place and link the boxes, and should have been removed before the utility went public.

      Like

      Reply ↓

      • That still leaves the “tree” diagram as a question. How do I read the different branches? What is the difference between the “series” lines as opposed the “parallel” connecting lines?
        Or, on the text version, how do I use the “From” (relating to a previous TG number) line?
        Finally, the tool provided me with one (or more) TGs with only one kit. How can that be a real TG?
        I am sure that I have other questions but, until those questions are answered, I don’t know what the other questions are? 🙂
        Thanks for your willingness to help all of us newbies!

        Like

  23. Jim – I remember reading that you’ve mapped over 90% of your segments to the contributing ancestors. I’m doing okay at the grandparent level, thanks to close relatives who have tested, but progress is slow after that since it is so difficult to triangulate with 5th-8th cousins. I would love to see a blog post on advanced techniques for finding MRCAs and solving segment matches with more distant cousins. Thanks,

    Like

    Reply ↓

    • Rich, I think the secret is in contacting every Match in a TG; and trying to get them to communicate with each other. None of us knows all the answers – there’s a good chance that some of our Matches in a TG could find the CA among them, without my input – perhaps they’d come up with evidence on the other side of my brick walls and/or NPEs that I didn’t realize. We need to use the synergy of a TG team.

      Liked by 1 person

      Reply ↓

  24. I am new to DNA and have tested at Ancestry.com and FTDNA. I greatly appreciate all this information. I intend to study it. I am 80 and my “forgetter” works over time so I need the help. I live in South riding VA just outside od DC and I see you do a lot of lectures in this area so I would be interesting in coming to one of your lectures.

    Like

    Reply ↓

Leave a comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.