Triangulation for Autosomal DNA

In our last article, Triangulation for Y DNA, we covered triangulation for the Y chromosome, how it works, and how it can help a genetic genealogist.

In this article, we’re going to cover triangulation for autosomal DNA.

Triangulation for autosomal DNA is kind of a chicken and egg thing.  The goal is to associate and identify specific DNA segments to specific ancestors.  The easiest way to do this, or to begin the process, is with known relatives.  This gets you started identifying “family segments.”  From that point, you can use the known family segments, along with some common sense tools, to identify other people that are related through those common ancestors.  Through those matches with other people, you can continue to break down your DNA into more and more granular family lines.  This is easiest to visualize thinking about your 4 grandparents.

Triangulation is easiest if you have parents or grandparents living, and you can test them.  Yes, all of them.  Their DNA will give your immediate pointers when you have matches to which side of the family you share with your matches.  If you can test your 4 grandparents, you immediately know which of those 4 lines someone who matches you descends through, because they will also match one, and hopefully only one, of your 4 grandparents.  However, for some of us, testing even one parents is simply not possible, so first, let’s look at some examples of triangulation without your parents DNA results.

I’m fortunate that one of my cousins has given a lot of focus to our Vannoy line.  Vannoy was the surname of my great-grandmother, Elizabeth Vannoy (1846-1918) who married Lazarus Estes (1845-1919).  The Vannoy line has a mystery we’ve been trying to solve for decades now called, “Who Was Elijah Vannoy’s Father?”.  Elijah was Elizabeth’s grandfather.  Your family probably has a similar mystery, and these tools hold the potential to answer those questions.  They also have the potential to introduce more questions.  But then again, isn’t that the way of genealogy?  For every ancestor we find, we get two more questions.

Several of the Vannoy cousins are interested in solving this mystery as well, so they have taken the autosomal Family Finder test at Family Tree DNA.

We know how they are related, and the men have all been proven to be Vannoy via Y-line testing.  By doing this, we’ve assured no undocumented adoptions, also known as NPEs (NonParental Events) in the Vannoy line.

We expect our cousins to match, and indeed they do.  This is my test result showing my three cousins who match me.

In my family mystery, “Who Was Elijah Vannoy’s Father?”, there are 4 candidates, all brothers who lived in Wilkes County, NC in the late 1700s.  Elijah was born in 1786.  We have the wives surnames.  Hickerson is our primary candidate surname, so I wanted to see everyone who matches me on my match list who also shows the Hickerson surname.  I enter that surname in the “ancestral surname” box, and click on “run report.”  The matches returned will all carry the Hickerson surname, which you can see by scrolling for the highlighted names. Turns out, it was only my Vannoy cousins – today – but tomorrow might be different.

Vannoy match 1

Now for the triangulation tool.

I want to see if these three people share common DNA not just with me, but with each other.  If we all share a common segment of DNA, then that confirms a common ancestor and attributes the DNA at that address on that chromosome to that specific ancestral family.  This is the fundamental concept on which triangulation is based.

In my case, the known ancestral family is Vannoy, not Hickerson, at least not yet, so let’s look at the Vannoy cousins as compared to me.

vannoy match 2

Each of the participants results are color coded.  On the page below, you can see that each matching segment of the chromosomes is colored.  It turns out that all of us share a fairly large segment on Chromosome 15.  So now we can attribute that segment to Elijah Vannoy, our oldest proven ancestor in that line.  You can also see some areas where one or two of my cousins match my DNA, but not all of us.  Those can also be attributed to Elijah Vannoy’s line since we share no other (known) common ancestors.

vannoy match 3

This cousin match is simple because the men share the same surname, but if this was 3 women with different surnames, the matching would still work.  The challenge of course would be to find the common ancestor.  In this case, if all 3 women had Elijah Vannoy in their tree, we could still tell that this segment of Chromosome 15 was attributed to the Vannoy family because they all matched me and matched each other as well on the same DNA segment.

Eliminating False Matches

Now let’s move to the “what ifs.”  When my kids were young, I just hated sentences that started with “what if.”

What if I have a fourth match, Jane, with unknown ancestry who matches me on these segments, but does not match any of my cousins?

To determine this you would also have to look at your cousin’s matches or ask Jane if she also matches those cousins.  Remember that half of your DNA is that of your mother and the other half is that of your father.  You will have people that match you, and potentially on the same segments as your known relatives match you, but are not related to both you and your relatives.  This means they are matching you on the other half of your DNA.  In this case, if Jane didn’t match my Vannoy cousins too on that same segment of chromosome 15, then we would know that Jane’s match would be from my mother’s side.

To illustrate this point, let’s move to my results at 23andMe.

Let’s use Family Inheritance Advanced to see an example of two people who match me on the same segment, but are from opposite sides of my family.  My cousins Stacy and Cheryl are from Dad’s and Mom’s side of the family, respectively.  We know they don’t share common ancestry, but look, they both match me on four of the same segments.

cheryl stacy match

How is this possible, you ask.  Remember, I have two halves of each chromosome, one from Mom and one from Dad.  It just so happens that Cheryl and Stacy both match me on the same segment, but they are actually matching two different sides of my chromosome.  For this reason, these are called HIRs, or Half Identical Regions.

Now let’s prove this to the doubting Thomas’s out there.

cheryl stacy match 2

Here is the comparison of Cheryl and Stacy directly to each other.  They do have one small matching segment, 6 cM, so on the small side.  But they don’t match each other on any of the segments where I match both of them.

If they did match each other and me on the same locations, it would mean that we three have common ancestry.

The fact that they match each other on one segment could also mean they have distant common ancestry, which could be from one of our common lines or a line that I don’t share with them, or it could mean they have an identical by state (IBS) segment, meaning they come from a common population someplace hundreds to thousands of years ago.

The real message here is that you can never, ever, assume.  We all know about assume, and if you do, it will.  In this case, assuming would have been easy if you didn’t delve into the big picture, because both of these family lines contain Millers from Ohio living in close proximity in the 1800s.  However these Miller lines have been proven not to be the same lines (via Yline testing) and therefore, any assumptions would have been incorrect, despite the suggestive location and in-common names. Furthermore, cousin Stacy’s Miller line married into her line after our common ancestor, so is not blood related to me.  But conclusions are easy to jump to, especially for excited or inexperienced genetic genealogists.  It’s tempting even for those of us who are fairly seasoned now, but after you’ve been burned a few times, you do learn some modicum of restraint!

So, what’s next?

Color your Chromosomes

In my article, “The Autosomal Me – the Holy Grail – Identifying Native Genealogy Lines,” I described in detail the process of downloading your DNA information from either 23andMe or Family Tree DNA and then utilizing that information in a spreadsheet to look at matches – not 3 or 4 matches at a time, but chromosome by chromosome.

In my case, I was fortunate to have my mother’s DNA results at Family Tree DNA before she passed away, and I was equally as fortunate that they were still viable for the Family Finder test.  Believe me, I held my breath.

Because I have her results, I can tell immediately if my matches are from her side or from my father’s side.  If the person matches both Mom and me, then it’s from her side.  See how easy triangulation is.

Let’s take a look at Chromosome 15 with all of those Vannoy matches on my spreadsheet and see what kind of information we can glean.

vannoy table 1

On my master spreadsheet, my Mother’s matches have been copied in and are color coded, but since none of these people match Mother, I have eliminated that aspect here to avoid unnecessary confusion.

The people identified as “Dad” mean that I know they are genealogically related on my father’s side.  People who match Mother genetically are labeled Mom.  There aren’t any on this segment of chromosome 15, in our example above.  The blank cells in that column, by inference, match Dad’s DNA, since they don’t match Mom.  When I confirm genealogically how we’re related, I’ll enter “Dad” in that column, but not until then.

I’d like to comment on information gleaned from the spreadsheet.  Every DNA segment has a story to tell.

Cousin Estes

First, Cousin Estes, with yellow highlighting, is one of my closest Estes relatives.  He is a third cousin on the Estes side and also descends from Lazarus Estes and Elizabeth Vannoy.  He matches me on the segment from 26 (million) to 58 (million). My Vannoy group of matches, shaded green, extend from 33 to 58, so this tells me that the area from 26 to 33 where I match Cousin Estes, and not any Vannoys, is probably from an Estes ancestor, and not the Vannoy line.

Unfortunately, I don’t have any other matches on this segment, so I can’t figure out which line it comes from, just yet.

The green areas are common between me, cousin Estes and the Vannoy cousins.  If we could find a Hickerson match on these same segments, we could then solve the family mystery AND attribute part of this DNA to the Hickerson line.  But so far, no dice.  This is why it’s important to continue to look and to reach out to people you match, especially those who don’t enter their family surnames or post a GEDCOM file.  The answer may be waiting for you.

The Insanity Factor

The pink segment labeled Cousin Younger is making me insane, so let me share some insanity with you.

The Younger line descends through the Estes line, significantly upstream. The Y DNA of Marcus Younger, who had 1 son who had 1 son, does not match the expected Younger DNA line in Halifax County, Va.  Cousin Younger’s only solid Y match also doesn’t match his expected family line, so we’re fish out of water on the Y-line.  Two undocumented adoption cases that match each other, but no one else.  Great, just great.  These are the things genetic genealogy nightmares are made of.

Mary Younger, daughter of Marcus Younger, married George Estes who fought in the Revolutionary War.  Their son John R. Estes married Nancy Ann Moore in Halifax County and they settled in Claiborne County, TN about 1820 where the Vannoy family is found as well, having migrated from Wilkes Co., NC.  John Y. Estes, son of John R. Estes had son Lazarus Estes who married Elizabeth Vannoy.  Here’s the generational progression:

  1. Marcus Younger – wife unknown, Y DNA doesn’t match Younger line
  2. Mary Younger married George Estes, Halifax Co., VA
  3. John R. Estes married Nancy Ann Moore, moved to Claiborne Co, TN
  4. John Y. Estes married Rutha Dodson
  5. Lazarus Estes married Elizabeth Vannoy
  6. George Estes married Ollie Bolton
  7. My father, William Sterling Estes

And of course, there’s a monkey-wrench, so let’s throw it in.  Marcus Younger’s grandson, ancestor of Cousin Younger, married a Moore woman in Halifax County, VA.  We believe we know who her parents are, but we’re not positive.  If they are who we believe, Y-line DNA tests say the 2 Moore families, living within sight of each other, aren’t the same Moore line….but they interact closely and my Moore line doesn’t match any Moores upstream anyplace.  So, we have another unknown ingredient in the soup.

So, from me, Marcus Younger is 7 generations upstream.  I should carry about 1.5% of his DNA.  I was pleased to see that my Younger cousin and I matched.

However, and this is a BIG however, the Vannoy line should not be related to the Younger line.  We know that both of these cousins are matching on my father’s side, not just because of the genealogy, but because neither matches my mother.  But they are somehow related, as Cousin Younger is matching the Vannoy group big as life on chromosome 15.  Could this be an IBS (identical by state) segment?  Yes, it’s small – but I’m not comfortable relegating it to IBS because it’s genealogically “inconvenient,” at least not yet.

So, something may well be wrong, amiss or unknown in the genealogy, either in Tennessee, which is doubtful as we have that fairly solidly nailed down, especially in recent generations, or in Virginia where there is at least one known disconnect and possibly two taking into consideration the Moore monkeywrench.  Still, the Vannoy family was not living in the same state as the Younger family and came from New Jersey to North Carolina, not from Virginia.  Maybe the connection is in one of the unknown wives lines.

So, you can see my reason for being perplexed.  One thing is sure.  DNA doesn’t lie.  It’s up to us to figure out the message it is conveying and which ancestor it is from.

Powerful Tools

I hope you can see what a powerful tool we have at our disposal.  Of course, it can reveal who your ancestors are, along with some surprises.  I don’t mind the surprises.  I view them as gifts from the ancestors.  It’s those crazy-making half-surprises that bother me.  I swear, the ancestors have a sense of humor.

118 thoughts on “Triangulation for Autosomal DNA

  1. Pingback: Ancestry Shared Matches Combined With New Ancestor Discoveries | DNAeXplained – Genetic Genealogy

  2. Pingback: James Crumley ( c1711 – 1764), Slave-Owning Quaker Moonshiner, 52 Ancestors #88 | DNAeXplained – Genetic Genealogy

  3. Pingback: Hannah Mercer (c1740-c1773), Died at 33, 52 Ancestors #89 | DNAeXplained – Genetic Genealogy

  4. Pingback: Ancestry Shakey Leaf Disappearing Matches: Now You See Them, Now You Don’t | DNAeXplained – Genetic Genealogy

  5. Pingback: Autosomal DNA Matching Confidence Spectrum | DNAeXplained – Genetic Genealogy

  6. Pingback: DNAeXplain Archives – Introductory DNA | DNAeXplained – Genetic Genealogy

  7. Pingback: Looking for and Contacting Birth Family Members | DNAeXplained – Genetic Genealogy

  8. how can you find out if your ancestor or ancestors were native American, which test is that is it this one? thank you , peggy l;ycans

  9. Thanks for all the information on your blog. As a newcomer to DNA genealogy, it has been most helpful.

    But this post raises an interesting question for me, when you showed the chromosome browser and your matches on C15. For C15 is one of my best matches too. I cannot find any of the people you mention in my FTDNA matches, yet when I lay your browser for C15 against mine it suggests that the three people you show match me and several of my matches in two segments – the first about 15 cM, and the second about 18-19 cM (just by eye).

    Here’s a graphic showing my browser (top) and your one in this post (bottom). I have two other matches similar to the green and about 8 other similar to the blue and orange.

    Are these matches at all significant? Thanks

    • I’m not positive I understand what you are asking, but I think you’re asking if because the colors are the same, if it’s is significant. The colors are always the same on the chromosome browser because they only have 5 spaces to fill. The only significance is if you match someone else on that same segment and share a common ancestor. My matches don’t compare to your matches and aren’t relevant unless we match the same person on the same segment and also match each other.

      • Sorry, I wasn’t clear in my thinking or in my writing. No, I’m quite aware that the colours are arbitrary. But the chromosome browser graphics show that you match a number of people on the same segments that I match a number of people. I can’t check whether we match each other, but I wondered whether it was common for people to match different people on the same segments.

        But now I think about it more, I know there must be millions of different matches on the same segments, and the fact that you and I have matches on the same segments of C15 is nothing more than a trivial coincidence. I am at a bit of a brick wall and I just got a little sign of (false) hope and I reacted too quickly. I’m sorry to waste your time, but at least I’ve got my thinking right now. Thanks.

  10. Pingback: We Match…But Are We Related? | DNAeXplained – Genetic Genealogy

  11. Pingback: The Ancestry 200 | DNAeXplained – Genetic Genealogy

  12. Pingback: Ethnicity Testing – A Conundrum | DNAeXplained – Genetic Genealogy

  13. I have a quibble (or maybe just a really dumb question) about the term “triangulation.” To my (plane geometry) mind a triangle has only three sides. Thus triangulation to a Most Recent Common Ancestor (at the apex of that triangle) would involve only two DNA testers (who form the base).

    It seems to me that what you describe should be called (tongue-twistingly) “tetrahedronation,” since it requires three testers to form the base of a 3-dimensional structure in order to locate the MRCA at its apex. I’m sure I’m missing something, and that my geometric, GPS-like, conceptualization is naive, but can you explain (probably for the umpteenth time) why a DNA match between only two testers, both of whom have well-researched family trees which show the same ancestor, isn’t good enough? Is the issue simply one of statistical probability, or is there some more fundamental difference, geometric or otherwise?

    • Good question. Here’s why. Let’s say that those two people who’ve researched the tree have an unknown second common line. They THINK their common line is the one they are researching, but if their other common line is a dead end line to one or both of them, they’ll never figure that out. If a third person from the line they believe they descend from doesn’t show up and match them, then they can’t be sure. A second scenario I’ve seen is that the wife was married twice. Both people descend from the wife, but not the second husband. But the trees were wrong. A third person who descends from another child of the second husband (especially if from his first wife) confirms the correct ancestor. A third person who descends only from the first husband will never match the pair of people who are matching through the wife.

      • I like to think of it as like map triangulation. Where two forest fire spotters on platforms on different mountain tops take an angle reading to a fire at a third point and figure out where it is.

      • Wow, Roberta, that was a quick reply. Thank you! If I understand you correctly, the difference between a 2-person (2-D triangulated) match and a 3-person (tetrahedronated) match is basically one of statistical likelihood that the identification of the MRCA is correct. Put another way: even with three matches there could still be a chance, albeit a smaller one, of a dead end scenario.

        For instance, let’s say AncestryDNA’s computer excitedly tells me that I am one of three members of a circle. I am confident about my research that identified my MRCA in this triangulated group, but the family trees compiled and posted by both of the other two circle members were simply copied from someone else’s, erroneous, data, as seems to be the case with public trees fairly often. In other words, having three (or more) testers forming the base of a tetrahedron (or a pyramid with four or more sides) simply improves the odds that a MCRA identification is correct, because it becomes progressively more unlikely that all of the other trees used to make a triangulation would be incorrect.

        And thus the difference between matches involving two persons or three persons is quantitative (statistically) rather than qualitative, right? And, pushing my GPS metaphor (the navigation acronym, not Genealogical Proof Standard :-)) harder, using three matches improves the chances of a correct identification somewhat like the way a GPS receiver’s localization precision depends upon the number of satellites that are in its field of view.

        Your DNAeXplained website is a truly outstanding resource. Thank you for all you and your colleagues do to help rookies like me deal with the complexities of this exciting, but often arcane, new field. I greatly admire your teaching style, and I love the allusion in your article on Autosomal DNA Matching Confidence Spectrum to the “lions, tigers, and bears” we beginners encounter as we venture into this frontier along a new yellow brick road.

  14. Pingback: Concepts – How Your Autosomal DNA Identifies Your Ancestors | DNAeXplained – Genetic Genealogy

  15. Roberta; I’ve got two matches that both match me and both match each other, but we don’t match all three on any one common segment. I can show how we are all related genealogically, and I do understand how by the distance we share (3rd – 4th cousins) that we might have unded up with different parts of the same people. If I understand the concept of triangulation, this is not a true triangulation. Am I missing something here?

  16. Pingback: Curtis Benjamin Lore (1856-1909), Devilishly Handsome Rogue, 52 Ancestors #113 | DNAeXplained – Genetic Genealogy

  17. Pingback: Concepts – Identical by…Descent, State, Population and Chance | DNAeXplained – Genetic Genealogy

  18. Pingback: Concepts – CentiMorgans, SNPs and Pickin’ Crab | DNAeXplained – Genetic Genealogy

  19. Pingback: Concepts – Parental Phasing | DNAeXplained – Genetic Genealogy

  20. Pingback: Jacob Lentz (1783-1870), Shipwrecked, 52 Ancestors #121 | DNAeXplained – Genetic Genealogy

  21. Pingback: Trying to Make Ancestors Out of NADs | DNAeXplained – Genetic Genealogy

  22. Pingback: Concepts – Managing Autosomal DNA Matches – Step 1 – Assigning Parental Sides | DNAeXplained – Genetic Genealogy

  23. Pingback: Margaret Lentz (1822-1903), The Seasons and the Sundays, 52 Ancestors #124 | DNAeXplained – Genetic Genealogy

  24. Pingback: John David Miller (1812-1902), Never In His Wildest Dreams, 52 Ancestors #125 | DNAeXplained – Genetic Genealogy

  25. Pingback: The Stages of Genetic Genealogy Addiction | DNAeXplained – Genetic Genealogy

  26. I got a question. I have a relative on gedmatch who apparently shares enough autosomal DNA to be considered not just a cousin but a full sibling. Unless my parents adopted out some unknown sibling of mine, I possess one full sibling & one female half sibling [the relative is male apparently]. 3523 versus the 2678 I get with my sibling, 1.0 generations versus 1.2 generations.

    Thing is, is this relative doesn’t share any chromosomes with myself or my sibling. Nor for that matter grabbing a bunch of random relatives and using gedmatch’s chromosome segment matrix anyone else.

    So is he a relative or some glitch?

  27. Pingback: Nine Autosomal Tools at Family Tree DNA | DNAeXplained – Genetic Genealogy

  28. Pingback: Concepts – Match Groups and Triangulation | DNAeXplained – Genetic Genealogy

  29. I’m working with a pool of a couple hundred people who descend from a common ancestor based on y-DNA. The paper trail suggests their common ancestor lived in about 1700 or maybe the 1600s. Many do not match at 7cm. I am attempting to triangulate and taking the cM level down to 3. At 3 to 5 cM, some of them triangulate. A number of the triangulations are clustered around two or three Chr. In addition, about half of the triangulated groups have two pairs of matches that share the exact same start and end on the cM.
    1) Are these triangulations simply the result of random numbers aligning? The relationships coming from the triangulations make sense.
    2) What is the significance of the clusters at certain Chrs? Do they help to validate the triangulations even though the matches share only a small Cm.
    3) What is the significance of two paired matches having the exact same start and end markers?

    Thank you.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your 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