15 May 2011

Triangulate to find more meaningful matches using both Family Tree DNA and 23andMe

Updated May 8, 2013 with new 23andMe website navigation and match figures:

The allure and potential of the latest generation of autosomal DNA testing is great for genealogists.  It instantly builds your family tree for you!  Uh, just joking.  But there are several real benefits of this type of DNA test:

  • it helps you confirm genealogical paper trails across ALL of your family lines (not just the paternal and maternal lines as yDNA and mtDNA tests do)
  • it helps you expand your family trees by finding "genetic cousins" and other relatives (for the sake of simplicity, I'll call these people Matches in this write-up)
  • it works especially well for the most recent five or so generations
  • both men and women can take the test
  • it only requires 6 pints of your blood (sorry, I couldn't resist.  No, the test doesn't involve blood, just spit)

Most of us who have taken Family Tree DNA's Family Finder test or 23andMe's DNA Relatives test have lots of Matches to work with.  Those Matches are typically lots of potential cousins that you may be able to connect with and share information in order to help each other both validate and expand your family trees.  So what's the problem?

Strangely, it's often that there are lots of Matches to work with.

I've tested with both companies, and I have 302 Family Finder Matches and 1,021 DNA Relatives Matches!  So, what's a decent strategy for narrowing this large number down to something more manageable?  Or at least for prioritizing the list so as to not get buyer's remorse after lots of non-responses and even more unfound connections.  And how can both company's test results be used together to find more Meaningful Matches?

Here's what I've done...
To achieve "genetic genealogy success" with over half a dozen different Matches and an equal number of ancestors so far in the last several months, here are the general steps I follow with both Family Finder and DNA Relatives:

1.  BE CLOSE.  Start with Matches shown as 4th Cousin or better.  After you've exhausted those, you can attempt 5th Cousin Matches (I've confirmed at least one family line with a Match listed as a 5th Cousin, but the connection generally gets more difficult to find and prove when you're dealing with more distant relationships).  Given that I have over 1,000 Matches on DNA Relatives I can narrow the list down to 124 who are listed as "3rd to 6th Cousins" or better in their Relationship range.

2.  BE INTERESTED.  Know that people on Family Tree DNA's Family Finder site are typically more universally interested in genealogy that those on 23andMe's DNA Relatives site (this is mainly due to the fact that many people use 23andMe for health purposes and the DNA Relatives test may have just come along with their purchased package).  But I've actually had better results with the DNA Relatives Matches so far.  Why is that?  I only make contact with Matches who list a decent number of surnames in their 23andMe profile.  It stands to reason that if someone has no or very few surnames listed in their profile, they 1) likely aren't very interested in genealogy, 2) haven't had much success finding their ancestors, or 3) in some other way may not be too helpful in the mutual quest for ancestors.

3.  BE SIMILAR.  Of the list of Matches you've narrowed down so far, look for people who have one or more matching surnames to surnames you know are in any of your direct bloodlines in the last five or six generations.  Make contact with the Match, and in the case of 23andMe, be sure to share genomes at the basic level.

4.  TRIANGULATE.  Here's the step that I've had the most success with:  Of those Matches you share common surnames with, find more than one person (Match) who matches you along the same section of the same chromosome.  It stands to reason that if you share a long segment along a given chromosome and you have a confirmed genealogical paper trail that proves a relationship with that person (Match 1), then if you share much of the same segment of the same chromosome with another person (Match 2) who also shares a surname in common with you and with Match 1...let's just say that often where there's smoke there's fire.  This step involves some minor spreadsheet action, but it's not difficult:

In Family Tree DNA's Family Finder, go to Chromosome Browser, click the name of the Match you're working on, then click the View this data in a table link on the right side of the page.  In that table, I tend to just look at Chromosomes with a matching length of around 10 or more CentiMorgans AND a 1,000 or more Matching SNP's.  There's no perfect number; you just want the "larger" numbers.  See example screen below of a row I would choose for this Match.  Hold tight, and I'll tell you what to do with this.

In 23andMe, go to My Results, Ancestry Tools, then Family Inheritance: Advanced.  Choose the name of the Match you're working on, then click the View in a table link.  It only shows you the longer-matching segments, so you already have the data you need for the next step.  See example screen below.

Now you're ready to triangulate!  To do this, I simply built a spreadsheet with the columns of data found in the tables above, plus a few others to help me keep track of my Matches.  See the screen shot of my spreadsheet below (with the names of my Matches partially hidden to preserve their privacy).  Click the image below for a larger version.  I'll explain the additional columns at right, and how I use them.

Here are a few quick explanations of the columns of the spreadsheet and how I use them:

  • I added a Source column to keep track of whether I found the Match on Family Tree DNA or on 23andMe
  • I added a Surnames in Common column at the far right to list surnames that the Matching person and I both stated were in our direct bloodlines in the most recent dozen or so generations.  I also added a Speculative MRCAs column and Confirmed MRCAs column to type in the name of the Most Recent Common Ancestor that my Match and I share.  I use the Speculative MRCAs column when my Match or I haven't found strong enough genealogical evidence to prove the link, and the Confirmed MRCAs column when we both already have strong documented evidence of which most recent ancestral couple we both relate back to.
  • Each time I find a new Match, I sort the spreadsheet by Ch (Chromosome) and Start Point (the start of the range along the chromosome which my Match and I share in common)
  • I color the rows where more than one Match and I have a large contiguous segment which all overlap.  (The different colors don't mean anything, they're just to highlight what's similar.)


Now here's the exciting part
.  Once I did the spreadsheet work above, I then look for at least one surname that is shared among the same-color-highlighted Matches.  For example, look at rows 25-28.  For three of them, we had Close and Smith as surnames in common.  I made further contact with those Matches, and we shared more of what we knew about the Close and Smith lines of our family trees by providing access to online family trees, documents, and so on.  For three of the four of those Matches, we were able to determine that we all descended from different children of my 4th great grandparents, William Rhesa Close and Belinda Varnum Smith!  (We haven't yet determined how the Match on Row 26 and I related)

Quite cool indeed, as this not only solidified and expanded the genealogy for that line of our families, but I've also met several new genealogy cousins who have been great to correspond with.

Well, this post grew to be a bit longer than I had envisioned.  Sorry about that; but I hope it is useful in highlighting a few ideas that perhaps you haven't tried yet.  Feel free to add comments and suggestions for everyone's benefit below.


p.s.  For another way I've used DNA testing -- this time Family Tree DNA's yDNA test -- read this article:  Unexpected Name-Change Brick Wall Busted!