Size Matters for Matching DNA Segments

The more large segments you share, the more closely related you are to a DNA match. To learn more about segments, read the new blog by one of my favorite DNA writers on the yahoo DNA-NEWBIE list, Jim Bartlett, called Segmentology.org
Brynne Ancestor MapHe explains why you share such large chunks of DNA with your closer relatives by showing how recombination works and how few cross-overs there usually are in a recent post at  http://segmentology.org/2015/05/24/segments-bottom-up/

While the total amount of shared DNA in centimorgans (cMs) is a fine indicator for your closest relatives out to second cousins, farther relationships can be more and more random in the amount of shared DNA. You can consult the charts and numbers at ISOGG here – http://www.isogg.org/wiki/Autosomal_DNA_statistics – to determine those closer relatives.

Although I like this new chart from Blaine Bettinger the best:

Shared DNA statistics from Blaine Bettinger

Shared DNA statistics from Blaine Bettinger, used by permission

Note that since you share about 25% with an aunt, uncle, grandparent, grandchild or half sibling the amount shared cannot tell those relatives apart. Sometimes the X can help distinguish those, see my post on that here.

Frankly the testing companies are not that accurate in predicting your more distant relationships, those past 2nd cousins. So when deciding which DNA matches to pursue, follow up on the ones with more than one large segment; those are your closer relatives, the ones whose trees may be easier to connect to yours.

42 thoughts on “Size Matters for Matching DNA Segments

  1. With my two successful matches, or matches with whom I found a connection, I know how unpredictable the size can be. And recently, after having a lot of cousins tested and two of them are 2nd cousins to each other and they share the largest segment 40.4cM, total 157cM, to me was a good indicator on what 2nd cousins should be sharing, more or less. Since I get pages and pages of 2nd cousins, if not predicted closer (FTDNA), it was good to go by this.

    My mother has had a 50cM largest segment match (total 122cM) and knew that this had to have been a close match. Took me over 7 months to figure that out (actually 11 straight days of non-stop research) to find out that my mother and this match are 1/2 third cousins. Pretty distant in a way, but that 50cM is pretty large. The next largest segment is 38cM.

    My mother gave my brother basically that same largest segment whereas I got only 14cM of that same exact segment. I actually share 28cM of that 2nd largest segment, but had I not got my mother DNA tested (I tested first) I wouldn’t have paid attention to that match, and actually I didn’t until almost a year after my mother tested.

    Definitely from 2nd cousins on it can be unpredictable.

  2. Kitty, a great post – and thanks for the plug. My true brother and I have such a mish-mash of DNA, that the Admixture/Ethnicity programs look like we are not related! DNA is random and very variable.

  3. Kitty – Great post. I’m so glad you like the chart! I’m still working on the Shared cM Project, there’s so much still to do, stay tuned!

    And I agree, Jim’s blog must be added to everyone’s must-read list!

  4. Hi Kitty,

    I am trying to understand the significance between “Total Segments” and “Largest Segments” on Gedmatch. If you have let’s say, 8 matching segments with someone, and the largest segment is maybe 4 or 5cM, and the total of all segments is maybe 17cM, is that of significance when trying to find a common ancestor down the line? I guess I’m confused with terminology because I keep reading in many places that matches over 10cM are pretty good ones, but I don’t know if this means 10cm on one segment alone or a total of multiple segments.

    • Reg,
      In your example the segments are too small for the relationship to likely be findable.
      If there is no endogamy in your tree then anyone who matches you on multiple segments of 7cM or larger is worth following up on. Segments smaller than that are often false, below 5cM they are false more often than true.
      Use total cM of 7cM and larger segments with the charts above which you can get on the one to one compare

  5. I am more or less the kit Admin for my adopted Japanese daughter in law. I noticed two of her matches, a mother and daughter have a lot of small segments. The mother daughter pair are probably 3c1r and 4th cousin. The daughter has the most looking at 3cm she has 130 cM total, other 4th cousins have maybe 30 to 40 cM some a little more. Does this mean anything? I was thinking maybe they are a little closer than they appear?

    • Alan –
      Small segments can be population specific. We tend to only look at matching segments greater than 7cM and total those. Also large segments greater than 20cM are indicative of a closer relationship not lots of little segments.
      Look at the numbers here after you total the larger segments.
      http://isogg.org/wiki/Autosomal_DNA_statistics
      Also if the daughter has matching segments that the mother does not, they would come from her father. Matching on both sides like that can make a relationship look closer than it is.

      • Hi MsKitty,

        Love all of your data, i am just getting my 23andme results back and trying to reach out to some of my “Closest” cousins haha. I am talking to one that i match two segments on the X chromosome at 22CM and 9CM i assume i should definitely pursue research there right? I mean we only share .45% DNA, but having two over 7CM on the same Chromosome matters right? SOrry, i am new

  6. i share546 centimorgans shared across 24 DNA segments with a match …
    trying to figure out the relationship ..she said its her moms side of the family and she is the half sister and her brother gave a baby girl up for adoption ??

    • Deanna – I am not sure what you are asking. That is on the low side, but the DNA match could be a half-niece relationship. If you are an adoptee or she is, then I recommend the DNAadoption.com web site plus join their mailing list at yahoo and ask questions there. Or DNA detectives on facebook.

  7. Hi Kitty,

    My Mom shares 89.8, 3 segments (chr 1,18,21) longest 74.4 (chr 1) with a male match and
    she also shares 187.2, 6 segments (chr 5,6,7,9,14), longest 78.4 (chr 9) and (chr 14 at 41.5 and 15.3) plus 26 “X” cMs with another male match. Any significance in 2 long segments within one chr?

    I’ve found the same MRCA in their trees but they don’t match each other on Ancestry or Gedmatch! It’s like they’re not related to each other (and not in common with my Mom) but I know they share this Hood line. Can you share any wisdom on this issue and both having long segment matches with Mom? Both are in their 60’s and Mom is in her 80’s…

    Mom’s 187 match thinks it’s possible his great Uncle or a 1/2 great Uncle may be my Mom’s great grandfather. Which means they would share a 2nd great grandparent. I’m having trouble with this because they share more DNA than 3rd cousin level? Thank you!

    • Debbie
      My apologies, I seem to have missed your comment. How long ago do they share the Hood line. A few 3rd cousins do not match at all (10%) and about half of 4th cousins do not match. See
      https://isogg.org/wiki/Cousin_statistics

      And that is a fine amount for 3rd cousins to share. Blaine’s latest chart shows average 79cM and range 0 – 189

      The amount of DNA shared past 2nd cousins gets more and more random …
      Kitty

  8. Hi Kitty, I am trying to find my Ashkenazi father. This is one of my closest matches from his side from Gedmatch. Any insight on what type of relationship this is with the endogamous component. Trying to decide if it is worth proceeding with, what do you think? Thank you so much for considering.

    Minimum threshold size to be included in total = 500 SNPs
    Mismatch-bunching Limit = 250 SNPs
    Minimum segment cM to be included in total = 7.0 cM
    Chr Start Location End Location Centimorgans (cM) SNPs
    1 242,426,244 245,689,563 8.9 634
    3 5,828,299 71,791,127 78.9 10,022
    6 6,813,583 10,288,551 7.1 588
    16 60,358,369 76,240,659 12.9 1,896
    Largest segment = 78.9 cM
    Total of segments > 7 cM = 107.8 cM
    4 matching segments
    Estimated number of generations to MRCA = 3.5
    397067 SNPs used for this comparison.

    • Definitely worth pursing that match, about a 3rd cousin, 2C1R, 3C1R, 4th …. the large segment is promising for a closer match.

      See Lara Diamond’s family statistics for an idea of what endogamy does to shared cMs in relationships past 2nd cousins:
      https://larasgenealogy.blogspot.com/2016/04/how-endogamy-looks-in-practice.html

      My new rule of thumb for DNA cousins past 2nd on the AJ side which meet the one segment > 20 and another > 10 criteria is to halve the total cMs and then look that number up on the chart …

    • Sorry Wendy, I seem to have missed your comment. a large segment like that is usually closer family 3rd cousin or closer but it is the total cM which we use to estimate the relationship as in the chart shown above.

  9. Hi, i cannot find the segment lengths from my testing company. I surprisingly matched to several members of a family I do not know. They range from less than 400 cM to around 175cM, over 12 but under 20 segments. Is there any way that this possibly is IBS? I did this for med info and ethnicity for fun, this matching was a total shock. Please advise.

    BTW, Do you do consults? Kind of…..bowled over.

  10. Thank you for your article. It has been one of the most helpful in trying to understand smaller segments and their possible significance or non-significance.
    I understand that if there are a small number of matching segments, there is little chance of relationship.
    If there is a segment that is over 7 cMs and it is accompanied by a small number of matching segments, it may or may not support the indication of relationship.
    My question is, because I have a number of matches where this is true, I will have a match with someone where the total of matching segments (using the 7 cM default) is over 25 cMs, even up to 60, and those matches are accompanied by a large number of segments (20 to 30) which are smaller giving a total number of shared cMs between 60 and 100.
    Do those smaller segments support the relationship?
    Do they indicate a closer relationship?
    Does paying attention to the chromosome where the segments are found help to understand? (For example, on the 8th chromosome there are 5 segments which add up to 12 cMs)?
    Thank you for taking the time to answer.
    George

  11. My advice is to ignore most of the smaller segments below 7cM and use the total of the larger segments with the chart to make a guesstimate of the relationship.

    Sometimes the 6s and even the 5s can be valid and triangulate with other relatives but for the most part they are false.

    Having many of the smaller segments will often occur when you are from the same population group so likely you have shared ancestors many hundreds of years ago, further back than the records. That is all a plethora of small segments might indicate. Not relevant to the closeness of your relationship.

  12. Hi
    I have tested with several DNA companies & my closest match on my Ashkenazi side is on MyHeritage. The match is:Shared DNA 1.7% (125.2 cM)
    Shared segments 4
    Largest segment 82.8 cM Is This a match worth contacting etc. I am an adoptee searching for my birthfather.

  13. I am reposting an issue of mine, to get an expert opinion (analysis) from someone on this Blog who maybe an expert on small segments. I’ve read there is a lot of debate on the topic and is usually negative. I’ve read ISOGG and Blaine Bettinger’s blog and others. There is an issue, where on a surname line came from the same small town as my same surname line came from. I really thought we were related as 4th to 5th cousins. Ancestry match results bottom’s out at 6 cM. With the background on my line, some of the matches with my surname, I would say the cM’s shared were really low in that my grandfather and my great-great grandfather did not pass much shared DNA down to me according to a chart, ‘DNA Detectives Autosomal Statistics Chart’ and other online statistics charts. I was wondering to myself, how would I match to 4th to 5th cousins, or have the same common 4x great grandparents.
    So I uploaded the results to GedMatch which you can lower the Centimorgans to 3. I compared myself to the other line with start to end as follows: 23,278,836 to 28,530,339, 4.5 cM, 900 SNPs on Chr.5 and 27,093,528 to 47,702,256, 5.0 cM, 1,220 SNPs on Chr. 16. I compared myself to the other line’s Aunt with start to end: 23,568,400 to 28,355,305, 4.2 cM, 823 SNPs on Chr. 5 and 27,273,899 to 47,576,695, 4.4 cM, 1,126 SNP’s on Chr.16.
    I checked my niece’s kit to the other line realizing that she would be a generation offset(more distant) and this dropped the Chr. 5 results. I compared her to the other line with start to end as follows: 27,729,894 to 47,573,693, 3.5 cM, 1,021 SNP’s. I compared her to the other line’s Aunt with start to end as follows: 27,729,894 to 47,573,693, 3.5 cM, 1,015 SNP’s on Chr.16
    All three of us have common, somewhat overlapping segments on Chr. 5 and 16 with somewhat strong SNP’s, with at least 4 cM not 3cM’s. Even my niece’s less than 4 cM’s results, hung in there with Chr.16 with rather large SNP’s. In some study on small segments that quote, “When utilizing small segments, I generally don’t drop the SNP threshold below 500”. SNPs are larger than 500.
    My ancestor did change his first name when he came to America on the boat, but there are documents such as parish/church records from here and in Germany. Why he changed his name, not sure why, but it could be a confirmation name which are not documented in parish records in Germany.
    Would you think that my line and the other line with the same surname are related (Identical by Descent)?

  14. This is way too complicated for a quick answer on my blog but yes this could be valid, especially with a large number of SNPs. Test more cousins on each line to try to confirm. Oldest generations best.

    Small matches in cM of 4-6 with high SNPs are possibly valid if they triangulate but beware, these can also be population segments. Look at who else matches you there.

  15. This seems high for a 2nd cousin. I know who it is, we thought we were 1/2 first cousins. Isn’t this more within a 1st cousin range? Thanks!

    Chr Start Location End Location Centimorgans (cM) SNPs
    1 59,215,941 120,075,978 59.4 6,288
    1 179,426,082 194,391,631 11.9 1,085
    2 42,188,689 74,825,589 33.0 3,734
    2 77,198,999 119,936,267 27.8 2,853
    3 38,411 28,993,092 51.8 4,120
    3 128,952,856 157,910,521 30.3 2,676
    3 185,884,816 196,259,734 26.1 1,422
    8 30,337,787 68,738,971 25.8 2,556
    9 5,686,021 33,398,019 43.0 4,039
    9 71,030,216 79,391,121 9.2 1,007
    11 10,373,830 107,128,834 91.8 9,180
    13 35,350,141 111,813,664 92.5 8,150
    15 67,889,745 98,108,585 52.3 3,753
    20 24,474,724 53,492,826 33.5 2,768
    22 14,884,399 24,620,588 25.9 1,091
    22 25,392,038 49,524,956 50.4 3,300
    Largest segment = 92.5 cM
    Total of segments > 7 cM = 664.6 cM
    16 matching segments
    Estimated number of generations to MRCA = 2.2

    • That looks fine for half first cousins, a little on the high side but DNA is pretty variable. It could also be full first cousins as you suggest. Another possibility is that the parents who are not expected to be the same are related. Any endogamy in that tree?

      You can confirm the full versus half by testing 2nd or 3rd cousins or other close relatives that you expect NOT to be related to both of you. If they are only related to the expected person, that would confirm the halfness. If they are related to both, but very unequally, it could be the related issue. So you may need to test even more people.

  16. Hi,

    My Mum has a match from another country with an SNP of 627 and 8.6 cM. It states around 6.8 Generations. Is this a positive match? This match would mean us. Thank you:)

  17. Awesome. Thanks so much. Its pretty exciting to me. My ancestors were traded, so having a match from the continent is very amazing. I understand it will be very distant though. I will be testing much older relatives to see if they have the same match too.

Leave a Reply

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