Using Genome Mate (Identify triangulated groups and confirm segments)

The previous article in this series is Using Genome Mate (Importing/Updating the email addresses of your matches from GEDmatch).

In following the steps that have been previously outlined, we now have imported a sizable amount of data from GEDmatch into Genome Mate. This post will show how to begin working with this data to identify groups of triangulated segments. The process of identifying triangulated groups can be very time consuming, and like genealogy itself, is something that is never ending. The concepts of identifying triangulated groups are fairly simple to grasp, but there is no shortcut to taking the necessary time to work through and analyze your DNA data. As one leading genetic genealogist, Dr. Blaine Bettinger (http://www.thegeneticgenealogist.com), summed it up so accurately, “Finding genetic matches is easy, but finding the common ancestor from whom we inherited a segment [of] DNA is very hard.”

There can be many factors that go into determining whether or not a DNA segment is indeed a valid match with another person. My goal with this post is to demonstrate the basic principals of triangulating DNA segments, and provide a general overview of the process, rather than overwhelm the reader with all the many small details that can quickly overwhelm a genetic genealogy newcomer. To this end, knowledgeable readers will recognize that I’ve simplified some of the concepts of DNA matching shown in this post, and that I’m not showing the GEDmatch triangulation tool. This is by design. We need to learn the basics before we can move beyond them into all the nuances of triangulation. In future posts, I will cover the more detailed concepts of DNA matching, or provide links to where these concepts are covered more in depth, for those wishing to continue with learning about all the fine points of genetic genealogy. I will also show how to use the GEDmatch triangulation tool in an upcoming post. First, let’s get a bit more acquainted with how Genome Mate displays our matches.

After having imported some GEDmatch data, the main Genome Mate screen will look similar to this screenshot. Unless you’ve changed the default, the segment data will be displayed as a graph, which facilitates quick visual identification of where overlapping segments occur. While we are viewing this screen, let’s look at three items that will be helpful as we move forward with identifying triangulated groups and confirming segments.

Genome Mate Triangulated Segments Post Background Screen Shot

The first item to take note of is the chromosome number dropdown box, as I’ve highlighted by the red rectangle. This dropdown box is used to step through all the chromosomes to see the matches you have on each chromosome. The chromosome selections are numbered 1 through 22, plus there is also an additional X chromosome selection.

Genome Mate Screen Shot 30

The next item to take note of is the Start and End columns for the segments you share. The start and end points listed on the graphical display of the Genome Mate screen are rounded to millions. For example, the segment I share with PA340556P1 Lawrence Stone actually has a start point of 11,244 and this is displayed as 0 (aero) in Genome Mate since the actual number is less than one million. The segment I share with PA340556P1 Lawrence Stone actually has an end point of 63,374,274 (displayed as 63). The segment I share with A242029 R. E. H. (administered…) actually has a start point of 12,171,406 (12) and an end point of 24,425,401 (24). The column to the right of the End column is the cMs column, which is commonly referred to as the length of the segment, as expressed in centimorgans (cMs).

Genome Mate Screen Shot 31

To see the full start point and end point for a segment, move your mouse cursor over the relative name or the segment bar, and a box will appear showing the specifics of the segment.

Genome Mate Screen Shot 33

The last item to make note of is the graphic display of your chromosome map for this chromosome. The upper bar is the paternal side of your ancestral line, and the lower bar is the maternal side of your maternal line. Both bars start out as gray, and as you confirm your matching segments, colored portions will be added to the bars (as shown in this example). How to confirm segments, and fill in your chromosome map, will be demonstrated in the next post following this one. Now, let’s get started with the process of identifying triangulated groups.

Genome Mate Screen Shot 32

As you step through each of your chromosomes in Genome Mate, using the chromosome number dropdown box, you will be looking for groups of segments that overlap vertically, as I’ve highlighted by the red rectangles.

Genome Mate Triangulated Segments Post Background Screen Shot (With Red Outline Boxes)

We need to check each of the segments that line up vertically, to make sure they match each other, in order for them to be considered a triangulated group. An important concept to remember is that each of us inherits a set of chromosomes from both our mother and our father. Therefore, it is not possible to identify whether a segment you share with a match is from your mother’s side of the family, or from your father’s side of the family, by just looking at the segment itself. The determination of which side of your ancestral line the match falls on is done by comparisons with other matches.

For each group of overlapping segments you find, you need to identify which of them also match each other. When they match each other it is referred to as in common with (ICW). You can verify whether two people are ICW by running a GEDmatch ‘One-to-one’ comparison between these two people. Check the ‘One-to-one’ comparison result to make sure the same chromosome and segment these two people are sharing with you are also shared between them. If so, there is a common ancestor, or set of ancestors, shared by the three of you. When three or more people share the same chromosome and segment, this is called a triangulated group (sometimes referred to as TG). Anyone else who also shares the same chromosome and segment with a member of the triangulated group will also share the same common ancestor, or set of ancestors, and can be added to the group.

Note that although we’ve identified a triangulated group, we still do not know yet whether the common ancestor, or set of ancestors, for this group falls on the maternal side or paternal side of the ancestral line. If one or both of your parents are alive, and agreeable to taking a DNA test, it will greatly assist you with determining which side of your ancestral line your matches fall on, and is something I would highly encourage you to do. Even without DNA tests of one or both of your parents, however, it is still possible to determine which side of your ancestral line a match falls on, although the process can be a bit more involved.

After you’ve identified a triangulated group, even though we have not yet identified which side of your ancestral line they match on, if someone else matches with you on the same chromosome and segment as the members of the triangulated group, and does not match another member of the group on the same chromosome and segment, we will know that they are matching with you on the opposite side of your ancestral line from the side the triangulated group matches on. Let’s work through an example.

In this example, we will be working with the three overlapping segments at the far right, which I’ve highlighted in red, and labeled A, B, and C for reference (since I’ve blurred out the GEDmatch kit numbers of all the matches shown here).

Genome Mate Triangulated Segments Post Background Screen Shot (With Example Red Outline Box)

I start by doing a GEDmatch ‘One-to-one’ comparison between my kit and kit A, making note of the Start Location and End Location. You’ll notice that the GEDmatch ‘One-to-one’ comparison tool displays the full start point and end point of matching segments, rather than rounding to millions as Genome Mate does in the main Genome Mate screen.

Genome Mate Screen Shot 24

I next do a GEDmatch ‘One-to-one’ comparison between kit A and kit B, and check the chromosome number, and Start Location and End Location, of any matching segments. In this case, kit A and kit B match each other on the same segment of chromosome 2 that I match kit A on. This is an indication of shared ancestry between the three of us, and forms a triangulation group. [Note: Keep in mind that, due to several factors, the start locations and stop locations reported by GEDmatch are approximate, and not absolute. As long as the start locations between two segments are reasonably close (within about four of each other when rounded to millions), and the stop locations for the two segments segments are reasonably close (within about four of each other when rounded to millions), you can be comfortable in saying the segments are a match.] With this example I’m using, the Start Location is 181 when comparing my kit to kit A and 180 when comparing kit A to kit B (a difference of three million when rounded), while the End Location is 202 when comparing my kit to kit A and 199 when comparing kit A to kit B (a difference of one million when rounded), so I would consider the segments to be a match since the difference in start locations between the two segments is less than four million when rounded and the difference in end locations between the two segments is less than four million when rounded.

Genome Mate Screen Shot 21

I then do a GEDmatch ‘One-to-one’ comparison between kit A and kit C. They do not match each other on any segments at all. This is an indication that kit C shares common ancestry with me on the opposite side of my family tree from the side I share common ancestry with kit A and kit B. Therefore, kit C is not part of the same triangulation group that kit A and kit B are a part of.

Genome Mate Screen Shot 22

As you work through your matching segments checking for triangulation groups, when you do a GEDmatch ‘One-to-one’ comparison between two kits, be sure to verify that any matching segments between them are the same segment and chromosome you match match each of them on. If I got the result shown below when I did the above GEDmatch ‘One-to-one’ comparison between kit A and kit C, this would be an indication that kit C shares common ancestry with me on the opposite side of my family tree from the side I share common ancestry with kit A and kit B. The segment that kit A and kit C share is on a different chromosome (chromosome 1) from the segment I share with each of them (chromosome 2). Also, in this case, even if the matching segment happened to be on chromosome 2, the segment has a dramatically different start location and stop location than the segment I share with each of them on chromosome 2, and this would again be an indication that kit C shares common ancestry with me on the opposite side of my family tree from the side I share common ancestry with kit A and kit B. Therefore, since kit C does not match kit A, I know that kit C is not part of the same triangulation group that kit A and kit B are a part of.

Genome Mate Screen Shot 23

Once you’ve identified a triangulated group, the next step is to contact each of the testers in the triangulated group (in this case, kit A and kit B) to ask them about their ancestral lines. We are trying to determine which ancestor(s) of ours is also an ancestor of each of them. Sometimes this is rather easy to determine, but the majority of time it is not immediately obvious. Contacting your matches, and working with them to determine who you share as common ancestors, is where the greatest amount of time is spent in genetic genealogy research. Keeping track of your contacts with matches, using good notes, is essential. Genome Mate provides several tools to help with this process, and these will be covered in depth in future posts, but for now know that you can enter notes for a match by clicking on their name or segment bar, which brings up a details screen with a box for research comments at the lower left. In the next post, we will use this details screen to confirm our triangulated segments and add them to our chromosome map.

Genome Mate Screen Shot 25

As I mentioned earlier in this post, there are many factors which can go into making sure a segment is a valid match. I will be covering these factors in upcoming posts. For the time being, in order to maximize your success rate in identifying the common ancestor(s) of a triangulated group, my suggestion is to limit yourself to working with segments that are 10 cMs or greater in length. My reason for suggesting this is that statistics indicate segments at least 10 cMs in length are 99% certain of being a valid match. Even by limiting yourself to these larger segments, you should still have plenty of segments to work through, and once you’ve identified the MRCA for these larger segments, they will be very helpful for identifying the MRCA of the smaller segments after we’ve covered the factors that help determine a smaller segment is truly a valid match.

In addition to limiting yourself to segments 10 cMs or greater in length, I would also suggest you skip working with any segments for which the ICW appears in the main Genome Mate screen as orange or yellowish/brownish in color (as shown in the screenshot below), unless it is a known close relative (second cousin or closer in relationship). The significance of these orange and yellowish/brownish indications will also be covered in an upcoming post.

Genome Mate Screen Shot 37

Congratulations, we now know how to work through our matching segments and identify triangulated groups. In the next post in this series, we will learn how to confirm our triangulated segments and add them to our chromosome map. [Note: When you have finished a Genome Mate session, it is a very good idea to get in the habit of always backing up your database file before moving on to something else and/or exiting the program. You can never have too many backups of your data to fall back on, if needed.]

If you experience any problems while following these steps, notice that a screen has changed dramatically from what I’ve shown in the screenshots, or have an easier/better way to do the steps shown, please let me know by commenting on this thread or sending me an email. Thank you.

The next article in this series is Using Genome Mate (Add confirmed segments to chromosome map).

This article last updated 31 Dec 2014

Advertisements

28 thoughts on “Using Genome Mate (Identify triangulated groups and confirm segments)

  1. Pingback: Using Genome Mate (Importing/Updating the email addresses of your matches from GEDmatch) | Adventures In Genealogy Research: No Stone Unturned/The Wright Stuff

  2. Pingback: Using Genome Mate (Add confirmed segments to chromosome map) | Adventures In Genealogy Research: No Stone Unturned/The Wright Stuff

    • Hi Roberta,

      The post on identifying triangulated groups and confirming segments was quite challenging to write, as I wanted to make sure the basics were covered without it being overwhelming, so I’ll be interested in any feedback you may have after reading it.

      Thanks,

      Dan

      • Hi Dan,
        The write up is very well done! GEDmatch is overloaded today so I’ll have to test drive this later but will let you know how it goes. Off to read your second post.

  3. Thanks for this post. I am pretty new to all this and I think I have a similar example.
    Do I conclude C is related thru my paternal side based on the following?
    -> A to B (myself to confirmed maternal 2nd cousin)
    Chr 3 / 127 – 140 / 12.6 / 2735
    Total of segments > 7 cM = 217.7 cM
    -> A to C (myself to unknown relation)
    Chr 3 / 123 – 139 / 17.0 / 3779
    Total of segments > 7 cM = 17.0 cM
    -> B to C (confirmed maternal 2nd cousin to unknown relation)
    Chr 3 No Match
    Total of segments > 7 cM = 0 cM

    • Hi amoonrover,

      Good work. Your conclusion looks sound, but just to make sure, run the one to one comparison of B to C one more time, with the SNP count minimum threshold to be considered a matching segment lowered to 300 and the Minimum segment cM size to be included in total lowered to 3. I have seen instances where the one to one comparison tool did not report a match when there really is one. This can happen if the SNP total of the matching segment falls just under the default level of 700 and/or if the segment length falls just under the default level of 7. If B and C still do not have a match on chromosome 3 between start position 123 and end position 140, then I would say you are safe in attributing the shared ancestry with C as being on the opposite side of your tree from B. As you get more matches on this same segment, and can run additional comparisons, it will help to further verify whether this conclusion is indeed correct. Some of these additional points are planned for the upcoming post where I will cover some of the finer nuances of verifying segments.

      Thanks for your question,

      Dan

      • Thanks for the quick reply!
        I just ran the 1to1 comparisons for “A to C” and “B to C” at 300snp and 3cM. I saw no matching segments on any chromosome.
        -ed

      • Hi Ed,

        I appreciate you checking at the lower threshold levels and letting me know what you found. Given everything you’ve checked, if it was in my database, I would feel very comfortable marking segment C with an ICW Group of P in this case.

        Best regards,

        Dan

  4. Thank you for a very useful article – this is exactly what is needed, a step by step guide to using Genome Mate. It is clearly an excellent piece of software, but despite having imported lots of data I was at a loss as to how to make use of the imported data. I am looking forward eagerly to your next article in the series.

    • Hi Cliona,

      I’m glad to know the article was helpful. After you’ve had a chance to triangulate and confirm some segments, please let me know if you have any suggestions for other topics you would like to see covered.

      Thanks for your comments,

      Dan

  5. Dan,
    I’m just starting to play with this a bit now that GEDmatch is working for me. I have to go fix dinner pretty soon but hope to get back at it this evening.

    Poking around in Genome Mate this afternoon did give me and idea of a suggestion for future topics: Cleanup! I have a bunch of duplications because the same people have been added from GEDmatch and FTDNA. The cM counts often differ of course but the start and end are the same since all of the FTDNA people were transfers from Ancestry and the GEDmatch are also only Ancestry. I tried doing a deletion for a duplication for my uncle a while back and it created a mess. I need an easy button for this!

  6. Pingback: Using GEDmatch (The Triangulation Tool) | Adventures In Genealogy Research: No Stone Unturned/The Wright Stuff

  7. Pingback: Steps to DNA Matching and Chromosome Mapping with GEDmatch and Genome Mate | Adventures In Genealogy Research: No Stone Unturned/The Wright Stuff

  8. Pingback: Using Genome Mate (Add confirmed segments to chromosome map) | Adventures In Genealogy Research: No Stone Unturned/The Wright Stuff

  9. Dan,

    Thank you for the informative blog. I am just starting to get my feet wet leveraging DNA analysis for my genealogical efforts and your help is invaluable.

    I have a question. Suppose I find a DNA match on chromosome 4 with two other people (call them Ann and Bob). The start and end locations show as exact. But when I compare Ann and Bob’s DNA on GEDmatch they have the same exact end location as me, but their start location is lower than my start location – say by 10 million cMs.

    Does this mean I can’t confirm a match for all three of us? Or are we still a match?

    Thanks for your help and guidance.

    -John McMahon

    • Hi John,

      Thanks for the kind words.

      With the example you’ve given, this would indeed be a match as long as the segment portion that all three of you share is at least 7 cM in length and at least 500 SNPs. If the segment portion that is shared by all three people is less than either of these two criteria, the match starts to become questionable (becoming more and more questionable the closer you get to 0 cM in length and/or 0 SNPs).

      I hope this is helpful and answers your question, but if it does not, feel free to ask followup questions.

      Best of luck with your research efforts,

      Dan

  10. Hmmm.

    So If two other people and I share a long enough segment – even if the start and end points differ – then it proves a match? Better example:

    On Chromosome 8:

    (numbers in cMs)
    Joe and my match start on 12 and end on 52
    Bob and my match start on 22 and end on 62
    Joe and Bob’s match start on 22 and end on 82

    (Assuming SNPs are all above 500)

    Even though there is some big overlap – since we share the same 30 cM segment (22 to 52) then we would be a match?

    Within Genome Mate I have been looking for areas where matches line up fairly flush on the left and right – but it sounds like my primary concern should be the length in cMs of where we share a segment?

    Is that accurate?

    Sorry for the newbie question – I’m trying to understand how I should be analyzing this data.

    Thanks Dan.

    • You have summed it up perfectly, John. The size of the portion of the segment that you and the other two people share is the primary concern, rather than the segments lining up flush on the left and/or right. Just as you point out, despite a situation where there is a large overlap, you identify the smallest portion of the segment shared by the three of you, and then check to make sure that portion is above the thresholds that are generally accepted to indicate a match (7 cM and 500 SNP). Great job!

      • I’m slowly getting it (thanks to your patience!)

        Is there a way to estimate how many generations to a MRCA based on shared cM length? Is this what sites like FTDNA and GEDmatch do? Can you trust their estimates?

        Thanks (yet again) Dan.

      • Glad the information is helping. You are asking some great questions.

        Estimating the generations to the MRCA based on segment length are exactly what the genetic genealogy testing sites (such as FTDNA and GEDmatch) are doing. While these estimates can be helpful, they are only general estimates at best and vary quite a bit. The science is just too new to be able to make these generation predictions with more precision. The same is true of the admixture (ethnicity) estimates. Over time, as the science continues to improve, these estimates will likely get much better.

      • Yet another follow-up question. 😉

        Say I have a proven a match on my mother’s side on a specific chromosome and segment – call the person Ann. And say there is another match within the same chromosome and segment – call that person Bob. If Ann and Bob don’t show a match on the same segment, does that mean I can prove that Bob is related to me on my father’s side?

        Thanks Dan.

      • In the case you’ve outlined, you can reasonably surmise that Bob is related on your father’s side, but I would want to find another match to your father and Bob, on that same segment and tracing back to the same MRCA(s), before I would be comfortable saying the connection is “proven.”

  11. I’m just getting started and have about 15 relatives tested on ftdna.com. My plan is to add all of their profiles, gedcom file and ftdna/dnagedcom files into GenomeMate. At this point I believe I can find triangulated groups without gedmatch; just using the data I have where A=B&C and B+C.
    After going through the ftdna data, my plan is to add matches from gedmatch.com for each of the profiles.
    Is this a reasonable approach? Gedmatch is down at this point in time, but I don’t want to add the data and then have to start over because of an order problem.
    Thanks,
    Jim

    • Hi Jim,

      Everything sounds fine with your approach, and you should not have to start over in doing as you’ve outlined. In fact, you’ve actually touched upon one of the really nice benefits of using Genome Mate (having access to the downloaded data when the GEDmatch and/or FTDNA sites are down).

      Best wishes,

      Dan

  12. Hi Dan,

    I have just started to use genomate. I am working on finding out if matches are related for triangulation.

    Here is the example. My great aunt is A. She match B, C, D, and E on chromosome 1. B, C, and E have their data loaded into gedmatch. D is on 23andme.

    I compared A to B, then B to C, then C to D. How do I compare D to E? Or should I in this case, compare C to E? So, far A, B, C, and D share the same ancestor.

    The results are A to B is 1 to 25; B to C is 9 to 18; C to D is 14 to 27.

    Len

    • Hi Len,

      It sounds like you are off to a great start with your genetic genealogy research. If D only has results available on 23andMe, and E does not, there is no way to directly do a comparison between the two. It sounds like you have already properly triangulated the segment that is shared in common between A, B C and D, however, so a comparison of C to E should be able to answer whether or not E is part of this same triangulation group. Nice work, and best of luck to you on your continued research efforts!

      Dan

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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s