Finding Distant Relatives with Autosomal DNA Testing

Maybe you tested your DNA to prove or disprove a genealogical theory. Or maybe you tested to check on your health risks or carrier status. But now you see all these possible 3rd to 5th cousins in your family finder or relative finder lists and you wonder if you are related and if you can find that relationship. Perhaps you contacted a few and had almost no responses.

DNARelativesSampleSYes you are probably related, but without both of you having a good paper trail you would be most lucky to actually find that relationship. It is likely to be further back than suggested if your ancestors were at all endogamous. Just living in the same location for a few hundred years can lead to much inadvertent intermarrying and more common DNA than degree of relationship would expect. So autosomal DNA testing is no genealogical shortcut. Some of the people you contacted already know this, so if they saw no common surnames or places on your profile they may have lost interest.

So be prepared before contacting those likely 3rd to 5th cousins.

  1. Have an easily readable pedigree chart in both PDF and online format (GEDmatch can do the latter, more on GEDmatch later)
  2. Another good tool is a list of about 12 generations of ancestors by place name. Much easier for a possible relative to scan.
  3. Do some reading on the basics. A nice short article on autosomal DNA is this one on about.com and it has some more links.
  4. Try to talk some close relatives into testing so you have more data to work with
  5. Last but not least, make a decision on how much time you want to devote to this project … warning it can be addictive

If you want to do the minimum, then scan the localities and surnames of these possible cousins and contact the ones with surnames or place names in common with you. Indicate in your message that this is the reason for contact. Include the URL for your pedigree or family site and offer to send the PDF files. Include your email address if you are using the 23andme messaging system. The more directed the contact, the more likely you will hear back. I recently went through and cancelled some of my early invitations and send new improved ones on the lines above and got four new shares.

You might enjoy reading this relative finding success story from CeCe Moore on her blog.

If genetic genealogy and your DNA relatedness really interest you then you want to start a master spreadsheet of the DNA segments that you match your possible relatives with. If one or both of your parents are tested, then create one for each parent as well. I did a post about creating a DNA match spreadsheet a while back. This post will discuss more of the whys and wherefores. As a reminder, on 23andme you can find overlapping segments by selecting My Results in the top menu then Ancestry Tools then Family Inheritence: Advanced. At Family Tree DNA you select up to five people to show in the chromosome brower and then go to that browser.

Part of the objective is to find overlapping segments and thus clues to where the relationship might be. Many people enjoy building a picture of which distant ancestors their DNA came from when they do find the relationships (or just get more relatives to test).

So the first question is how large a segment is worthy of being included on this sheet.

The unit of measure is either a centimorgan (cM) or number of SNPs. Both those terms are linked to the ISOGG wiki articles explaining them.

Many genetic genealogists suggest that a one segment match of less than 10 cMs will not have a common ancestor in the last 200 years and thus hard to find the relationship. Plus matches smaller than 7cM may not even be IBD (identical by descent) since the smaller the match the more likely that it is not real, but rather IBS (identical by state). Remember you have two copies of each chromosome and the testing cannot say which side each allele came from.For example, if you are AA GG AA CC you will match me when I am AG AG AG CT but there is no guarantee that my A G A C came from a single side (paternal or maternal) so we need a much longer match to have some confidence that we are related.
There is a good explanation of this IBS vs IBD concept here: http://dna-footprints.com/203/the-abcs-of-dna-ibd-vs-ibs/

There are a number of charts on DNA relatedness, the ISOGG website has several, I like the one on their  IBD wiki page because I find the number of segments most useful. The family tree DNA FAQ has some very useful snippets also.

Personally on my spreadsheet I show all matches of about 6cM or more for any non Ashkenazi relative with three or more matching segments. I show single segment matches of 7cM or more. But those are just my choices, not necessarily real relatives. Since I have tested my father and brother, I can assume a smaller match is IBD when it is passed on (often a bit smaller) from Dad to one of us since our mother is from a completely different population group. My post on common relatives can be far back shows some of those smaller matches where the MRCA (most recent common ancestor) has been found, usually well before 1800 some before 1700.

The second question is where to find the matching segment data.

  • On family tree DNA you can go to family finder and matches, click  Chromosome Browser and painstakingly go through all your matches. Or you can assign the more interesting ones as possible relatives and then they are easier to find in that function. But there is a better way, use the tools at http://www.dnagedcom.com to download all your chromosome matches, then sort the file first by match size and discard all matches below your chosen threshold. Then create two more files from that file, one sorted by name and the other sorted by chromosome and start point. The second file is the basis for your master sheet
  • On 23andme, you can get a master list of overlapping segments from ancestry labs > ancestry finder. Wait for the page to load and then scroll down to the big button that says: Download yourname‘s Ancestry Finder Matches. However the problem is that many of these are listed as anonymous even when you are already sharing with them. So you need to share genomes with the anonymous ones to put a name to a segment. You send those invitations out via relative finder.  The tools at at http://www.dnagedcom.com will download all the matching segments from your shares. Once you have a combined master list make two more files, one sorted by name and the other sorted by chromosome and start point. The second file is the basis for your master sheet

When you find an overlapping segment with more than one possible relative be sure to check if they match each other too!

I have a number of cases where the same rounded off segment at 23andme which matches me is not a match between two relatives because they are related on different sides, one paternal and one maternal. To check at 23andme use the chromosome browser function and compare your two matches. At family finder, use the in common with function to see if one match is in common with the other one. Then they match each other. Of course that is no guarantee that they don’t match somewhere else instead so ask them if they match each other there or (best) compare them at GEDmatch. Alternately if you have other family members tested at family tree DNA you may be able to look at their match with these two (see a later post I did on that technique).

If you have uploaded your data to GEDmatch you can use the GEDmatch site to look at specific matching segments of interest to see if there are other possible matches. That is the “Show results that match on a given chromosome segment” function.  Then you can check if your possibles match each other as well by comparing those two id numbers using the “Compare your Autosomal FTDNA or 23andMe result with one other result in our database” function. The start and ending segments are shown as well as the cMs and SNPs.

Since my mother and father are from such different population groups, I now only do my mother’s side in my own spreadsheet and keep a separate spreadsheet for Dad. My current focus is Dad’s Norwegian ancestry (with a sprinkling of Finnish and British see his chromosome ancestry picture in my post about ancestry tools at 23andme) since records in Norway are great and very much online so we have a long paper trail. Plus we have solved one brick wall already with DNA!

If you want to go even deeper with your spreadsheet approach, I recommend joining the DNA_NEWBIE group via the ISOGG website and looking through their files on methodology as well as joining the genealogy-DNA group at rootsweb and reading all back posts by Tim Janzen.

More another day.

I have updated this post a little and added it to the DNA testing menu. Other posts to read as follow up are Jim Bartlett’s post on using spreadsheets and my page on the overlapping segment mapper, if you want to make a pretty picture from the CSV of your overlaps from your 40 closest DNA relatives.

17 thoughts on “Finding Distant Relatives with Autosomal DNA Testing

  1. Hi Kitty, I’m glad I found your blog. This post was helpful and very well written. It seems like you’ve learned a lot from this experience. I see that you have had both of your parents tested. Did you find any overlapping chromosome segments that both your parents shared just by looking at the chromosome data? – Thanks

    • Thank you Ginger, I am spending way too much time on this new hobby!! But no I only have my Dad tested as my mother passed away long before I took thus up.

      There is a function at GEDmatch that lets you check the relatedness of your parents from just your own DNA and it confirms that my parents are not related. There are 3 spots, two less than 1cM and one only a touch larger. Also my Dad’s parents from different parts of Norway were not closely related, one 2.5cM segment

      There are also more of these type utilities created by David Pike at his site – http://www.math.mun.ca/~dapike/FF23utils/

  2. In your post above, you wrote: “When you find an overlapping segment with more than one possible relative be sure to check if they match each other too! … To check at 23andme use the chromosome browser function and compare your two matches.”

    Did you really mean “chromosome browser” here? If so, could you provide a little more detail on how to do this?

    Thanks.

    • By chromosome browser I mean the function where you compare DNA segments that can be found by going to ancestry labs and then family inheritance advanced.

      23andme let’s you compare the DNA of anyone you share with against up to three other sharing people there

      • Ah, that makes more sense. When you said “chromosome brower,” I was thinking you meant the “Browse Raw Data” function (a drop down item under “Account”), and I was wondering if there was some feature in that item that I had somehow missed. Thanks for the clarification.

  3. mskitty this question is for you. I am a newbie with DNA but have some very basic knowledge of DNA only in the most general terms unfortunately. I am stalled at many of my lines due to them either being common names in NYC or being Irish, and did DNA testing in hopes of some enlightenment re my Irish lines (wishful thinking I know LOL) My question is this. I am descended from the vary first families in the US, which was a “closed population” for all intensive purposes. Those families intermarried over and over and I have TONS of matches that the only common denominator I can find is back at the 9 GGP level or so give or take a generation. I am related to some of the same people in several different ways due to all this intermarrying. Could this explain many matches at the 4th -6th cousin level (and a few at the 2nd -4th cousin level)??? Folks keep telling me it is impossible, autosomal only goes back 6-7 gens, but it seems to me that with all this intermarrying certain genes would be more prevalent. It is driving me crazy as people keep telling me this is impossible and I must have another relationship with these people at a lower level…but I sure cannot find it. Of course being stalled in most of my lines (Irish for the most part) does not help. I have thrown the book at my lines, trying everything I can think of and DNA is a last resort for me to try to make some headway. I am hoping someone else may have more info than I and I may find it via DNA. Anyway I would appreciate your opinion re these matches showing up for me, when all I can determine is that they come from these well known and well established lines back in the 1600-1700 time period in New England. If you are familiar with those folks we are talking the Munsons, Baldwins, Gunns, Northrups, Howland’s, Camps etc …huge families and tons of intermarrying. Thanks for your help and opinion. Joyce

  4. Yes the intermarrying will show many of your double 10th cousins as much closer. I have a triple 10th/11th cousin in Norway that shows as a 3rd-5th.

    Colonial ancestry is particularly known to do this. Ashkenazi DNA is even more prone to this effect, as are all very endogenous groups.

    Large families plus the randomness of DNA inheritance mean that you can certainly inherit DNA from as far as 10 generations back. It is not expected but with intermarriage it becomes more likely.

    Read this post for some more information on this:
    http://blog.kittycooper.com/2012/12/common-ancestors-can-be-very-far-back/

  5. I did find that after I posted the question. Thank you for clarifying. I have had person after person (even at GEDMATCH admin) tell me this was not possible except on very rare occasions, but I find many of my matches back to these folks…I have same issue on my Italian side LOL and I think this may be more common in earlier centuries than many people realize, esp with rural populations. My Italians married into same families over and over and my 1 x GGM was a cousin to her husband my 1 x GGF….It sure does ruin my day when I am trying to find my Irish guys and all these other matches confuse the heck out of things, esp when I run across people who have not gotten as far back as I have in the New England lines…on the good side I can point them in the right direction, unfortunately that only helps them, not me LOL . Thx so much for the answer. Lots of these folks often have Irish in lines as well. I read somewhere that one way “new” families to US tried to be accepted in US was to marry into old families…not sure how true that is, they took their secrets to the grave 🙂 Joyce

    • Thank you so much for your post! I have the same issue with my family. Intermarriage is making my research very challenging. It is nice to know I’m not missing something or getting confused just by the pure complexity of the subject and that my family is just a beast 😉

    • Oh, my, I read Joyce’s question from 2014 and mskitty’s reply, and yes, I will tell you I inadvertantly discovered this. My paternal ancestry goes back to 1635 and the founding of Hartford Connecticut in this country. Many many many intermarrying lines cross crossing. One fifth cousin in Canada and I discovered we have so many crossings that the software showing relationships gives us about twelve (fifth, sixth once removed, and so on). Since my ‘families’ relocated to Nova Scotia while they were still good British citizens, and since there were three successive waves of immigration prior to the American Revolution (Ulster Scots, Colonial Americans and the Yorkshire migration), we have three different chunks of shared ancestors, one going back to Leicestershire England in the fifteenth century, one going back to the Ulster Scots who were settled in the north of Ireland by King James and another from Yorkshire, where families in one Riding tended to marry within the confines of that Riding. Once my data was up on 23 and me, I discovered someone predicted to be a fourth cousin that I knew from the above researches could only be a sixth cousin as we shared fifth great grandparents. It made me look more closely at the shared relatives and any names listed, and I found precisely what mskitty explains.

      It’s enough to drive a person to drink!

  6. this is an excellent Blog, thanks for posting all the helpful insights you have learned. 23andme should hire you to provide guidance to their customers interested in genealogy.

    • Sorry, I seem to have missed your question William. Upload your data to GEDmatch.com and compare there. I will email you as well.

  7. Ancestry.Com indicated to me and another that we may be 1st cousins. After reviewing we cant seem to find such a direct link. Then we got in contact and found out that my possible 1st cousin was adopted and only knows who her biological mother is. She did have a first name and location from where he was from. Well the first name matches my father and he spent most of his life in Chicago. Although hard to pin down if he was in Chicago at the time. Since he passed away in October before we found out. Trying to use GEDMatch.Com to get more confirmation. Summary on GEDMATCH.Com is:

    Largest segment = 120.3 cM
    Total of segments > 7 cM = 1,824.4 cM
    37 matching segments
    Estimated number of generations to MRCA = 1.5

    Is this leading to conclusion we are half-siblings?

    • Yes you are half siblings. I will email you so you can send me the kit numbers to confirm.

      You might also test some paternal side relatives

Leave a Reply

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