Using GEDmatch to see where you match

I think the switch from Ancestry.com DNA test results, where your tree gets searched for you, making using DNA with genealogy easy – to GEDmatch where you have to figure out how to use the data yourself, is quite difficult. So this post is an attempt to help my cousins who have tested at ancestry and uploaded to GEDmatch. It might also help others new to GEDmatch who want to look at where they match a [possible] cousin by walking through that process.

GEDmatch Sample One to One

Sample from the GEDmatch one-to-one comparison

  • First make sure that you understand current DNA basics (click here for my page on that). Genetics have advanced greatly since my high school biology class and perhaps since yours too.
  • Next realize that the raw data from your test is only a small part of your genome, a sampling. It is the SNPs that are currently considered the most interesting. They represent the most likely spots where we are different from each other. If a contiguous sequence of those SNPs is the same in two people for about 10 centimorgans (cMs) or more then they are expected to share a common ancestor. With a match of 7-10 cMs it is likely but not a sure thing. There is a good article in the ISOGG wiki on the likelihood of a match at different segment sizes.
  • In order to see where your DNA matches someone else’s, you need your kit number and his kit number. Then you can use the one-to-one comparison to see on which chromosome(s) you match each other. Your kit number shows on your GEDmatch homepage. You can find the kit numbers of other possible relatives in the one-to-many display or perhaps your new cousin has sent you his kit number.
  • I recommend that you keep a spreadsheet with the information on your matches, sorted by chromosome and start point, so you can see who else a new match might match. I have a number of posts on this blog about using spreadsheets and a template in my downloads area. Many people like to use the genomemate tool to organize their data.

The image above is from a recent new match to my Dad uploaded from ancestry.com. The blue rectangle shows where there is a DNA match. The numbers in the box are what I cut and paste into my master spreadsheet for Dad.

Looking at my spreadsheet I see about 15 other folk matching on that segment, none of whom I have found a known ancestor with, but wait! Further up the sheet is a 37 cM match that continues down to this one. It is with my known 3rd cousin Katy so that puts this match on my WOLD line if he also matches her there. It is called triangulation when you have a three-way match. So next I have to do a one-to-one comparison of my new match with cousin Katy:

3rd cousin Katy versus new match

3rd cousin Katy versus new match

It’s a match! Now I will compare him to some of my other cousins on the WOLD line to see if he matches any of them elsewhere. I actually have a few distant cousins descended from the ancestors of my WOLD gg-grandparents so I might be able to further narrow down the possible common ancestors. None of them match this specific segment but a double 6th cousin on gg-grandma Anna’s line matches the segment just before it, that Katy has also, and segments do like to stick together.

When I am done with my analysis I will send Dad’s new match an email that includes a link to this article and to my online family trees and I will offer to look at his tree. In my experience on my Norwegian lines, a single segment match of this size will be for an ancestor in the late 1600s or early 1700s. Closer relatives will usually have more segments and larger ones.

Hopefully going through this process with me has been helpful. By the way I found cousin Katy with DNA and blogged about it of course.

For more of the fun things you can do at GEDmatch, like look at ancestry composition or use the chromosome browser, read some of my older posts about that site or view the slideshow I made for a GEDmatch presentation. There is also a manual for using GEDmatch in my downloads area.

Last but not least, I have a page here with answers to the most frequently asked questions on the DNA_NEWBIE list, which can help you with the terminology and methodology of genetic genealogy.

53 thoughts on “Using GEDmatch to see where you match

Click here to add your thoughts at the end of the comments
  1. Pingback: WikiWeek In Review: 6 October 2014 | WikiTree Blog

  2. Hi Kitty, I’m very new to genealogy and genetic genealogy. I am male. I have a kit in GEDmatch. I am waiting for Y-DNA results (trying to fill gaps on my paternal line with those); they aren’t available for a few months. I am trying to see how much utility I can get from just my atDNA and X-DNA results for now; I know my haplogroups for paternal and maternal, but don’t see how they help necessarily with my current objective. I am searching for a MRCA in an enormous family tree shared with me by a woman I match with in the range of 40 to 50 cM. I have no family tree to contribute on my paternal line. The X-DNA comparison on GEDmatch shows no match between us. My mental model of how X-DNA works may be incorrect. I am trying to narrow down whether I am related on her father’s side of the tree or mother’s of the tree and am wondering if our “no match on X” results can do that. My current thinking is the answer is Yes, but I may be wrong. Rationale: a woman gets an X from her Dad and an X from her Mom. I do not match this particular woman on the X-DNA comparison per GEDmatch. Therefore, I cannot be related to her father’s side of her family; else, there would be some match between my X-DNA and his X-DNA (which she inherited from the only chromosome he could give her). I must be related to her mother’s side. However, the X-DNA results alone cannot determine whether it is on this woman’s maternal grandfather’s side vs. her maternal grandmother’s side. Since the woman and I have no match on X, that must mean that the X chromosome she received from her mother is not the one I would have matched on. Nonetheless, the GEDmatch comparison of two kits on X will help me to narrow my search to this woman’s mother’s side of the tree as described above. Your thoughts on this line of reasoning are greatly appreciated. I’m ready for the bad news that my knowledge of how X works and can be used for tree-pruning is faulty, but I hope it is not! 🙂
    Thanks very much!
    -J. Michael

    • Hey J. Michael.

      I’m searching for my unknown father as well. Y testing has not given me any real clues just yet. But autosomal results has put me in the ballpark of comparing which male first cousin of a specific arm of a great great grandparent two of my DNA matches have in common.

      If right, I’m close to pinning enough evidence on being the direct male descendants on one of the original policemen who tracked down notorious bushrangers in colonial Australia. And I see how I was drawn to actual tracking as a scout in the infantry in my day…. not 100%, but in an ironic way, gives me something to keep me going on my search.

  3. J. Michael –
    So are you saying you do not know who your Dad is? Have you checked out the methodology at http://DNAadoption.com ?
    No you cannot make any assumptions based on the lack of X since this is not a close enough relative (About a 3rd cousin) and the X can be strange, see my articles http://blog.kittycooper.com/category/dna-genealogy/x-chromosome/
    However since you have no X from your father, if you are male, there is some likelyhood that this woman is related to you on your Dad’s side, but you knew that already.
    The best way to narrow this down is to get more of her cousins, aunts, uncles, tested, then compare your autosomal results to theirs.

    • Absence of an X-chromosome match proves nothing. Presence of a match only proves that that particular match occurs through a limited number of MRCAs. Theoretically, a male only inherits X from his mother, but a female inherits from both sides, so your mother’s X that you inherited, unchanged, is a mixture (not necessarily 50/50) that came from her father (unchanged from his mother) and her mother. Of course, her mother’s X is a mixture as well.

      • well put Byron. I would add the proviso that X can reach way back in time.

        So a smallish (less than about 8 cM) X match may not even be from the same path as the autosomal match. I have several cases of this on my tree. X matches to people whose known ancestor for the autosomal match is not on a possible X inheritance pathway. Of course they could also be IBC (false matches by coincidence).

  4. I still dont understand. I keep trying to understand what I’m reading but its like trying to read French books for an hour, and I dont understand because I dont speak or read French, only HS English level on top of it. I did a one to one and got pig latin results – not a clue what it means: Largest segment = 117.7 cM, Total of segments > 7 cM = 1332.7 cM, Estimated number of generations to MRCA = 1.7

  5. Yes like any field of study this requires a new language. It’s not as bad as philosophy though.

    Looking through my more basic posts I can that I do not define the concept of a matching segment anywhere which is the simplified term for an HIR
    http://www.isogg.org/wiki/HIR

    I will keep looking and if I do not find anything for you on this language I will write another blog post

    Have you looked at Kelly Wheaton’s lessons? Lessons 5 through 9 are about autosomal DNA
    https://sites.google.com/site/wheatonsurname/beginners-guide-to-genetic-genealogy/lesson-5-introduction-to-atdna

    Or Emily’s book recommended in my basics section?
    http://blog.kittycooper.com/dna-testing/dna-basics/

    My 2nd cousin is just getting into this so I need to write a post for hm anyway

  6. Hi Kitty. Could these results mean these are half-sisters, shared father?
    Chr Start Location End Location Centimorgans (cM) SNPs
    1 9,864,705 85,189,384 93.7 18,281
    1 106,669,295 247,169,190 146.9 28,784
    2 8,674 242,669,396 263.7 53,708
    3 2,643,594 39,339,606 56.1 10,441
    3 136,511,021 199,310,226 79.7 12,756
    4 61,566 5,179,542 7.8 1,005
    4 5,196,707 63,820,936 71.6 11,845
    4 63,833,261 155,006,540 77.1 16,309
    4 177,449,129 191,109,772 34.1 3,910
    5 131,812,292 178,551,313 66.9 11,613
    6 3,905,887 46,429,048 63.9 16,317
    8 6,401,356 146,255,887 154.3 30,802
    9 36,587 137,529,561 161.3 29,822
    10 12,808,796 135,297,961 143.9 31,496
    11 188,510 19,692,563 35.6 6,235
    11 103,047,628 134,436,845 50.6 8,842
    12 61,880 99,892,548 115.4 23,880
    13 17,956,717 109,134,842 118.7 24,312
    14 46,528,028 104,197,399 77.9 14,899
    15 66,267,941 95,747,735 48.9 7,601
    16 79,848,106 88,668,978 26.2 3,935
    17 45,698,385 75,513,707 53.8 7,404
    18 102,535 6,749,818 23.2 2,177
    18 69,604,024 71,823,892 7.2 822
    19 211,912 59,262,591 98.6 12,148
    20 973,258 62,374,274 109.9 16,722
    21 41,048,559 46,338,647 14.4 1,890
    22 16,050,224 39,479,811 48.7 6,009
    Largest segment = 263.7 cM
    Total of segments > 7 cM = 2,249.9 cM
    Estimated number of generations to MRCA = 1.3

  7. Since Ancestry.com already tells you how you are likely matched (sibling, 3rd cousin, 4th cousin, etc.) and gedmatch already tells you who you match with and how many cMs, what is it about the painstaking process of manually making matches that is more helpful?

    I can understand the aspect of seeing where on the family line they might be related to you, but isn’t that also automated? For example, if you go to ‘Find matches that match both kits’ on gedmatch, your cousin Katy should have just popped up there when you put in his kit number. I suppose I can see how looking at their chromosomes might tell you whether he’s related to her father or her mother, if you have their info, but I’m trying to figure out why you’d need to go to a spreadsheet and manually match when it does it for you.

    I’m certain I’m missing something, I’m just not sure what it is.

  8. Tarissa,
    What you are doing is fine and most people are content with it.
    However some of us are curious WHICH bit of DNA we have from which ancestor and wish to map our chromosome. See this post for a beautiful example:
    http://blog.kittycooper.com/2014/09/using-the-chromosome-mapper-to-make-a-four-generation-inheritance-picture/
    Others of us like to prove questionable lines from the paper trail with triangulation. My favorite is how we proved that Kristine’s WALD family were the same as my WOLD family.
    http://blog.kittycooper.com/2014/02/more-cousins-dna-tests-are-in-one-proveing-a-brick-wall-solved//
    Others are tracking down a specific trait in our DNA as “citizen scientists” – The men in my family often have a rare color blindness that is supposedly on chromosome 7 but I am busy proving it requires a specific bit of X in my family.
    So it all depends on how keen your interest in this stuff is as to how far you take it and whether segment information is useful. For adoptees it is invaluable in figuring out what family they fit into.
    Kitty

    • I am just beginning this process so much of it is new to me. My wife and her parents did the AncestryDNA test. When the results came back, her mother showed as her parent, but her father did not show as a match. Is this possible? Is it common?

      • Eljaykid,
        It’s called an NPE (non paternal event) and it happens about 10% of the time. Sometimes the husband could not get the wife pregnant so a donor, sometimes there are other back stories… This is a delicate problem.
        Do check with customer support first that there has not been a mistake. Your wife will want to ask her mother about this when she is ready. I will email you some more information.

        • Thank you for the reply mskitty. This is quite a shock and I don’t think it’s a mistake because his DNA kit matched his brother, so we’ve got some new exploring to do.

          Thanks again.

  9. Kitty,

    I am totally a newbie at this; I have done my DNA; I have uploaded it to Genmatch. I have found the first 2 matches and am in talks with the 2nd one, but the first close hit was 2.1.
    I did a little more research and found the girl that represents that match of 2.1 – she is much younger than I – almost my daughter’s age. Is it possible with that close of a match to chromosome 19 that her mother or father could be a lost sibling?
    I was given away at birth; I have found several sisters and 1 brother; one of my sisters and as well as me was given away at very young ages and we suspect there may have been another. Am I off base here with my assumption and if so what does the chromosome 19 (almost an exact match between her and I) mean.
    Thank you in advance Terri

    • Terri –
      That is a great match and could easily be a child of the lost sibling. Use the one to one compare between your two kits and see where the total cMs are with the numbers on this page to find the likely relationships: http://www.isogg.org/wiki/Autosomal_DNA_statistics

      As to the full match on chromosome 19, 19-22 are the smallest chromosomes and can easily be passed intact for more than one generation so no special significance other than a likely close relationship.

  10. I am also new and don’t understand the GEDmatch. I am adopted and looking for my father’s side. Will my 23 and me or Ancestry info give me any leads at all through GEDmatch? Do I have to have a biological father’s side male for that? I know not one person on his side and don’t have a name either. Many of the names look the same from both sites and I am sure they are maternal. Any ideas? Thank you, Bobby Sue.

    • Bobby Sue –
      Go to the website DNAadoption and have a look at their methodology. Yes DNA can solve this but it might take much time and patience or you could get lucky right away.
      Get maternal side family to test so you can separate which matches are from your mother’s side and which are not (your mother would be best of course).
      Be sure to test yourself at all three main companies, you can upload your ancestry raw data to Family Tree DNA for a mere $39 and be in a third pool.
      Be patient with yourself, it takes time to learn how to use all the tools at GEDmatch. I have a manual here in downloads and a slideshow at http://slides.com/kittycooper/gedmatch#/
      kitty

  11. Greetings mskitty,
    I’m trying to confirm if a young lady who showed up as a close relative\first cousin on Ancestry.com is truly my half sister.

    Here are some Gedmatch stats:

    Largest segment = 129.4 cM
    Total of segments > 7 cM = 1,875.7 cM
    Estimated number of generations to MRCA = 1.5

    From my research the 1,875.7 cM seems to suggest we are half siblings but I wanted to get your expert opinion 🙂

    Any help you can give me would be greatly appreciated.
    -Art

  12. Hello,

    I recently got my DNA results back on Ancestry and just uploaded them to GEDmatch. On Ancestry, I found a 1st-2nd cousin match on my father’s side. I don’t know my father, and he is not a match on my moms DNA. I have been in contact with him, and it seems he could be a match from his moms side, but we can’t seem to figure out how. My dad was adopted and he doesn’t know his birth family.

    Here are our one-on-one comparison on GEDmatch.
    Largest segment = 93.9 cM
    Total of segments > 7 cM = 1,171.9 cM
    Estimated number of generations to MRCA = 1.8

    What do you think?

    Thank you, Kim

  13. I think you are closely related, you could even be half siblings.
    Please either go to DNAadoption.com and post this on their mailing list with more details or try DNA detectives on Facebook

  14. Hi Kitty –

    Since almost all DNA testing is done by the Big Three (Ancestry.com, 23andMe, Family Tree DNA), and DNA results are matched by the latters, is there much to be gained by testing at other sites? Other than perhaps cross-testing Ancestry.com results to 23andMe if a person chose to test at only one company and not the other?

    I see a proliferation of matching sites:

    GEDmatch.com
    http://www.ysearch.org/ (FTDNA)
    http://www.mitosearch.org/ (FTDNA)
    https://dna.land/
    http://dnaadoption.com/
    https://dnagedcom.com/

    Am I paranoid, or are we seeing a covert and surreptitious attempt at amassing DNA, for ulterior purposes?

  15. Kitty,
    I have a client that has a “surprise” 25% match, 1824cm with an MRCA of 1.5. My question is that IF this is an aunt/niece situation, will the X Chromosome results be different depending on if it is a Paternal Aunt or Maternal Aunt?

    Thank you,
    Tabby

  16. Kitty,
    Yes, the half-sister scenario is a possibility and I have been exploring that option but I wanted to make sure I understood the full aunt/niece X Chromosome relationship. We are having more people in her family tested to narrow this down a bit. I was just wondering if, while looking at the X match, there was a way to tell which side the aunt/niece scenario would fall on. When you say that they should share 50% or 25% (if it is an aunt/niece relationship), which number should I look at to get that % on just the X Chromosome?

    Tabby

    Thank you,
    Tabby

    • Tabby – 50% and 25% are just statistical predictions; it will not necessarily be so at all. But if they match for the entire X then they almost certainly share a father.
      Best to have both kits at GEDmatch and compare X there. Then you can see visually how much they share and get the numbers to calculate percentages but nothing is sure with the X. It has been known not to recombine at all even from a mother.
      Please read the article whose URL I listed for a better understanding of X and also perhaps some of my other X articles

  17. Kitty,

    A little update. I did as you suggested and performed an X comparison and received the result ” No shared X-DNA segments found ” between myself and my possible half sister.

    As a reminder our Gedmatch Numbers are:

    Here are some Gedmatch stats:

    Largest segment = 129.4 cM
    Total of segments > 7 cM = 1,875.7 cM
    Estimated number of generations to MRCA = 1.5

    I recently had our DNA samples submitted for testing by another company for siblingship and the results are as follows:

    The probability of relatedness as half siblings is 98.2% as compared to an untested, unrelated random individual. The half sibling index is 56.19. The likelihood that the individuals are related as half siblings is 56.19 to 1.

    I’d be curious to get your take on their findings. Do you think their results exclude the Uncle\Niece or Aunt\Nephew possibility given our centimorgan numbers?

    Thank you!
    Art

    • It is extremely hard to tell half siblings from aunt/nephew. The main difference I have found is that half siblings will share more segments larger than 100 cM usually 4 or 5 while the aunt/nephew will often have only one or two. Plus half siblings will often have one or two fully matching chromosomes which is rarer in the aunt/nephew

      Sometimes X can help. If you share no X then you almost certainly do NOT share a mother (but in rare cases she could have given you different X)

      Send me the kit numbers via my contact form and I will take a quick look.

  18. Ms. Kitty!

    I have read through these blog posts but everything is still a little confusing… I was adopted and am searching for family. There is so much to look for but I would like to know what exact numbers or columns I need to look at. A very simple breakdown would be much help!

    • It is not simple, sorry

      Go to DNAadoption.com and look at their methodology

      Close relatives will share larger segments and ranges of total cM for a match can be found at the isogg wiki or on Blaine Bettinger’s blog (click his name in the cloud of tags on the right here)

  19. Hi Kitty,
    I have a match on ancestry that is 933 cms over 37 segments. Although it says cousin, I’m guessing it may be a half uncle or great uncle. The guy is the same age as my mother was. I’m searching for my father and he matches this side. He is not responding and I have found him on facebook but he is not responding on there also. I don’t believe its cause they don’t want to, I just think its cause they(his wife is the administrator) are elderly and don’t check these things often.
    Do you have any advise on the relationship or best way to contact them.
    Laurie

    • Your message may be in his other box so patience is advised. Some strategies. Check his friend list at Facebook, anyone you know there? If so try contacting them for help reaching him.
      Other ideas … Google his username from ancestry. Sometimes an email address will turn up on a forum. Else try sending an email to his username at aol, gmail, yahoo, hotmail – people tend to use the same usernames over and over again.
      Are you a member of DNA detectives at facebook? If not join and ask them for ideas.
      Good luck!

      • Thanks Kitty, I know his name and I have tried messaging so many people on Facebook, his wife, his children, even grandchildren with no responses. I found a phone number posted but it was out of service. I didn’t find anything from searching his name on Ancestry… At some point I gotta think maybe they are avoiding me. 🙂 I am a member of DNA Detectives and love the site. I’ll keep trying, there are still more options.
        Thanks!

  20. Some times they rather not get connected which can be super disappointing. My biological father told my half sister that myself and my brother both died of sids. (She half by mother’s side) His brother told me that he never mentioned me or my brother to anyone on that side of the family.

    I’m luckly my mother’s side of the family has been so welcoming and helpful. But sadly that isn’t always the case. Some times they just don’t know or want to be contacted

  21. Help me please. I am an adoptee with no family tree information. Apparently I uploaded my DNA 2 times and the only 4 digit match in the Total cM column seems to be my other DNA file. The next closest is all the way down to 191.2 and a Largest cM of 41.1. From there it goes down to all 2 digit numbers. This is also the next closest Gen match at 3.1 while my own other file is a 1.0, and the rest of the matches go back more then 4.1. Is it safe to assume that there are no relations that are anywhere close to me? I am trying to find a birth parent or living sibling so I am not really interested in researching long distant relations.

  22. I’m curious as to how MRCAs reflect grandchildren of half brothers. My grandmother matched up with someone on GEDMatch who I suspect is her second cousin once removed. They have an MRCA of 3.2, which means they should share a great grandfather (once removed if I’m reading it right), but I’m wondering if that’s still an accurate estimate if my grandmother’s grandfather and her match’s great grandfather were half brothers rather than full blooded.

    • The children of half brothers are half 1st cousins and will share about half the amount of first cousins so around 430cM.
      I always look at the total cMs and check for large segments if it is a close family match. That 3.2 from GEDmatch is just an estimate, albeit a good one, and it is showing about a 2nd cousin or 2C1R. The more distant the relationship, the more the shared cM can vary.
      A 2C1R averages 129cM but can be as high as 325 cM or as low as 0 according to this chart which I always refer to
      http://thegeneticgenealogist.com/wp-content/uploads/2016/06/SharedcMProject.png
      As well as the relationship statistics article at the ISOGG wiki

      • Thank you for getting back to me. My grandmother and her match share roughly 179 cM, which seems a little high if they’re half second cousins one removed. If I go by the max cM and halve it the reading should be at about 162, but the number I’m getting is higher than that. Maybe things will make more sense when my great aunt gets her results.

        • Yes it often takes additional testers to sort these relationships out. After close family, DNA inheritance gets more and more random that’s why I like that chart which gives a range.

Leave a Reply to J. Michael Cancel reply

Your email address will not be published. Required fields are marked *