Concepts – Match Groups and Triangulation

Today, we’re going to talk about the concepts of autosomal DNA and the differences between:

  • Match groups
  • Mathematical triangulation
  • Genealogical triangulation

Match Groups

At Family Tree DNA, when you download your chromosome matching results, meaning your complete spreadsheet, then sorting into chromosome order (sort by column, end location, start location, then chromosome) your spreadsheet will look like this.

MG1

Of course, your spreadsheet will be a lot longer and will continue with additional matches on chromosome 1, then chromosome 2, etc.

In the example above, we see that this is just one match group, meaning that the segments for all individuals overlap, which indicates that they match me. In fact, I copied the first match group on my spreadsheet to use in this example.

A match group is a group of people that match YOU on the same segment of your chromosome.  You will have many match groups.

Each of these people matches me on chromosome 1 beginning at location 72,017 for some distance. The shortest match is Calvin and he matches on only 3.47 cM, which tells me that I have other matches with Calvin, because this segment is too short to have made it over the match threshold by itself. So I know this is just one of at least two segment matches to Calvin.

I don’t know Calvin, so I don’t know which side of my tree Calvin falls on or how we are related.

The next several matches’ segments are about the same size, 11 or 12 cM, with the final segment being significantly larger, 26.82 cM. If you need a refresher, I wrote a concepts article about centiMorgans and SNPs.

There are two really important things to remember about match groups.

  1. Match groups means that the people who are on this list match YOU. It does NOT means that these people match each other. In fact, if you recall, you have two sides to each chromosome, one from Dad and one from Mom, so it’s very likely that you have matches from Mom and matches from Dad, intermixed, in this and every match group. Without additional information, you have no way to discern who matches you on which side.
  2. Do not be deceived by thinking that the beginning or ending location, or both, is indicative of matching on one side or the other, or that the people who share beginning and/or ending locations match each other too. If I were to fall into this trap, I would PRESUME (and that’s a very dangerous word in DNA matching) that Cheryl through Rutha, inclusive, match each other and are from the same parental side – and I will tell you right now – they aren’t. So just don’t go there because it will trip you up sure as shootin’.

So, the people in a match group match YOU. Don’t read anything more into these matches at this point.

However, let’s move on to mathematical triangulation because there’s more that we can discover.

Mathematical Triangulation

I’ve gone and snuck a new term in on you haven’t I – mathematical triangulation. Sorry folks.

We have talked about autosomal triangulation several times before. You can read about triangulation here.

In a nutshell, triangulation requires three things of all people in a triangulation group:

  • They all match you on the same segment (math), which you know because they are all on your match list
  • They all match each other on the same segment (math), which you may or may not be able to discern by utilizing various tools
  • They have a common ancestor or ancestral line (genealogy), which you may or may not be able to discern through traditional genealogy

Triangulation is two parts math and one part genealogy. Don’t let that math word frighten you, because the math isn’t the hard part as it can be done by sorting a spreadsheet or by vendor tools, but the genealogy has to be done by you.

Mathematical triangulation divides your matches into two groups, one from your mother’s side and one from your father’s side.

Let’s step through this process and see how it works.

On any group of people who match you on a particular segment of a chromosome, when you have enough people, your matches will form two groups, plus possibly an outlier or two.

Why?

Because you have matches from both your father’s side and your mother’s side, assuming enough people have tested.

Let’s look at an example.

If you have 3 matches, it’s possible that all 3 matches are from your mother’s side.  However, as more people test and match you, eventually, you will have two groups of people form, one from each parent’s sides.

How do you know who is in each group?

Good question – and this is what defines a triangulation group versus a match group.

In a triangulation group, all of the people in the group MUST ALSO MATCH EACH OTHER on the same segment.  Yes, I’m shouting because if you forget this, you’re toast!

How do you figure out if they match each other?

In my case, I have access to the kits of the people colored peach below, because I paid for their testing, or put another way, they tested to do me a favor.

MG2

The kit in blue is managed by a cousin with whom I have a many-years long relationship, so while I don’t have direct access to this kit, I do have a great working relationship with the person involved.

So, I can sign in and I can see who matches whom. If you don’t have access to any of the kits, you can look at the ICW (in common with) list, which tells you which of these people also match each other, but the ICW list does not tell you if they match on the same segments, just that they match.  It’s not triangulation, but it is information and if the entire group matches each other utilizing the ICW tool, that’s fairly indicative that you do indeed have a mathematical triangulation group – but it’s not 100%.  You can read about the ICW tool and how to use it in this Nine Autosomal Tools at Family Tree DNA article.

So, let’s see what the two groups look like after I see who matches whom by looking at each person’s matches on this segment of chromosome 1.

MG3

I was able to check each of Cheryl, Don and Rex’s kits and they all match each other along with Jim, so we know that all 4 of these people in the green group all match each other plus me.

I was able to check Lazarus’s kit directly and Amos’s kit through my cousin, and I was also to verify that they match each other and they both also match Rutha, so I know that this purple group all matches each other plus me.

So what happened to Calvin – the uncolored row in the middle?

First, I can’t check Calvin’s kit directly, but the fact that his segments do mathematically overlap BUT he doesn’t match all of the people in either group, in fact, he doesn’t match any of the people in either group whose kits I could check, tells me that his results are zigzagging back and forth between my mother’s and my father’s DNA, from side to side.  I also verified his non-matching status to my matches utilizing the ICW tool.

This is called an identical by chance match. In essence, Calvin doesn’t match either side so it’s what is known as a false positive match – looks to be real but isn’t upon further inspection.  Remember I said that your matches would form two groups that match each other – plus a few outlier?  Calvin is an outlier.

Now, we have all of our matches sorted into two groups, a mother’s group and a father’s group, plus one who doesn’t match either group. The question is which ancestors do these matches come from and which side is mother’s and which side is father’s.

Now it’s time to add the genealogy portion as the third piece of the triangulation pie.  Sometimes, when you don’t have the ability to do mathematical triangulation, per se, by comparing individual kits, you can achieve mathematical triangulation by utilizing genealogical triangulation – so these two actually go hand in hand.

While genealogical triangulation can achieve mathematical triangulation, by organizing people into matching sides, mathematical triangulation cannot replace genealogical triangulation.

But wait, there is actually a shortcut I can take, and it is a way to begin adding genealogy immediately and easily.

Genealogical Triangulation

There are multiple ways to perform the final step in triangulation which takes your matches from mathematically matched groups to genealogically triangulated groups with ancestors or ancestral lines assigned. In fact, sometimes the genealogy will actually be what helps you with the mathematical grouping if you don’t have access to kits to check who your matches match.

Genealogy Triangulation Method 1 – Adding a Parent or Parents

My mother has tested too. In my spreadsheet, I have added her matches into my master spreadsheet so I can easily see who matches both me and my mother. A match to both of us tells me immediately which side the match is on.

MG4

My mother’s matches are colored pink. You can see immediately that Mom and I both match Cheryl, Don, Rex and Jim, so those matches are assigned to my mother’s side.

Genealogical Triangulation Method 2 – Using Known Individuals and Identifying Common Ancestors or Ancestral Lines

As it turns out, I already know who Cheryl, Don and Rex are, so I know that they match on my mother’s side, even without my mother’s DNA test. Cheryl and Don are siblings who are my mother’s first cousins, and Rex is a second cousin. All of these people match both me and Mom on the same segments, which means that these matches come from my mother’s side.  It also means that I received this entire segment intact from my mother, without being divided.  Utilizing close relatives to sort matches into groups is exactly why we encourage everyone to test as many known relatives as you can convince to test, except children of relatives who have already tested, because their children only received a part of the parents’ DNA.

Furthermore, it also means that Jim, whose genealogy I don’t know, is from the same line because he matches Cheryl, Don, Rex and mother as well.

That does not means that Jim necessarily shares the same most recent common ancestors (MRCA) as Cheryl, Don, Rex and mother.

Even if I didn’t have my mother’s information, knowing her relationship to Cheryl, Don and Rex along with mathematical triangulation is enough to assign these relatives and people they all match to my mother’s side.  Even if I don’t have access to their accounts, I still know that they are closely related to my mother, there are multiple of them, and they all match me on the same segment, so I can group these people together, if nothing more.

Margaret Lentz chart

The common ancestors of Barbara (mother), Don and Cheryl are Evaline Miller and Hiram Ferverda.

The common ancestors of Barbara, Don, Cheryl and Rex are Margaret Lentz and John David Miller.

So while these individuals do share a common ancestor, and we can identify who it is, the most recent common ancestor is different between Rex and Cheryl, Don, and Barbara.

We don’t know who Jim is, but I can pretty much tell you that he isn’t a descendant of Evaline Miller and Hiram Ferverda, unless he is through an unknown child. I can be pretty certain that he’s not a direct descendant of Margaret Lentz and John David Miller either.

I can also tell you with equal conviction that he IS descended from either the Margaret Lentz or John David Miller line, because he matches all 4 cousins who descend from that couple – Cheryl, Don, Barbara and Rex – on the same reasonably large segment. So while we can’t identify his common ancestor with the group, at least not yet, we can say with certainly that he descends from either these common ancestors or an ancestor to these ancestors.

Now, if Jim also matched to William Lentz, above, which he doesn’t, but let’s say he did – we would then know which side of the Margaret Lentz and John David Miller line Jim represented. The Lentz line, of course.

This would also tell us that if Jim matched the Lentz line, that the DNA he shares with Barbara, Rex, Don and Cheryl was from Margaret Lentz, so descended from her parents, Jacob Lentz and Fredericka Reuhle. Of course, we don’t know that today – but all it takes is the right “new match” whose genealogy is proven and we can then attribute the individual segments to specific ancestors.

Let’s add mother’s lines into the mathematically matched chart.

MG5

As you can see, just as expected, Mom matches all of the same people that match each other, along with me, in the green match group, which is now a triangulation group because we know which side the match is on – Mom’s.

To add more definition to the triangulation group, we need genealogical information about the people in the group so that we know who their common ancestors are, or their common line. Fortunately, we have that.

You can also see that Mom matches Lisa as well, but neither Cheryl, Don, nor Rex match Lisa, so Lisa must be from Mom’s mother’s side AND I didn’t inherit that DNA from Mom because I don’t match Lisa either. That’s good information to know through deductive reasoning. It’s also possible of course that Lisa’s match to Mom is IBC, identical by chance, but at almost 14 cM, that’s rather unlikely.

I’ve updated the “side” column with what we’ve learned.

Lastly, given that I do know the genealogy of many of these people, I’ve added that information into additional columns on my spreadsheet, along with the fact that these segments do in fact triangulate. Please note that you can click to see a larger image.

MG6

Now, if you’ve just caught the words “these segments do in fact triangulate” and you’re about to ask if each matching segment needs to be triangulated individually – the answer would be yes. You may share multiple ancestors with someone, on both sides of your family. In fact, even worse you can share multiple ancestors on each side of your family. Endogamy will do that to you.

We’ll pause a minute here for the groaning to subside.

One last comment is that when I infer a side, like with Lisa who does not match on Mom’s father’s side on this segments, I don’t assign the side, I just make a note because we really CAN’T say that Lisa matches on my Mom’s mother’s side, because she might be a false positive match.

Also, in the case of Rutha, we know she descends from either one of these common ancestors or an ancestor of an ancestor, so I simply note the group she triangulates with for further reference.  That information about Rutha will come in useful when I work with other match groups that she is a member of – trying to make them into triangulation groups as well.

Genealogical Triangulation Method 3 – Phased Family Matches

You can also check for your phased Family Matches on your match page at Family Tree DNA to see if any of those individuals who match you on your spreadsheet are already assigned to the maternal or paternal sides based on phased matches with qualifying relatives. You can see on the page below that indeed, on my mother’s kit, Cheryl is assigned to her parental bucket, being her third closest match.

MG7

However, you CAN’T assume that because a match doesn’t have a maternal or paternal icon assigned that they aren’t descended from a particular side of your family.

Family Tree DNA only assigns high confidence phased matches so that you can depend on those results.

Remember, each segment needs to be individually triangulated, and the Family Matching algorithm that assigns maternal and paternal icons has a higher threshold and other internal requirements that may cause a parental icon NOT to be displayed when the match IS from that side of the family.  This is not a bug but a design element that assures that only highly confident matches are parentally assigned.

So you can use the parental icon as a great tool to assign a genealogical maternal or paternal side, but you still need to do due diligence in terms of working each segment to identify with whom it triangulates and the common ancestors.

Genealogical Triangulation Method 4 – Don’t Forget the X

While utilizing this trick won’t get you all the way to triangulation, it will help in several cases, at least by assigning parental “sides” to some matches – for males only.  Yes, ladies, I know, it’s not fair.

Because males inherit an X chromosome only from their mother, and a Y from their father, any match that is labeled as an X match:

  • Has to have come from a male’s maternal line if the segment is a valid match.
  • Has to have a segment on chromosomes 1-22 that is over the matching threshold for an X match to be reported.

Aside from that, the X is subject to the same segment size considerations of all other chromosomes and segments.

Any X match of a reasonable size, meaning one that is less likely to be identical by chance, had to have come from a male’s mother’s line – so a match to a male with a reasonably large X segment is an indication of a maternal line match for him, at least on that segment.  For two matching men, an X match has to be a maternal line match for both men.  However, keep in mind that I have seen X segments that match on completely different lines than autosomal matches.

When comparing the X chromosome, in non-endogamous populations, I would certainly note matches over 3cM. I would not assign a maternal “side” unless the match was 7cM or over and 500 SNPs or more.  Lastly, in endogamous populations, I would be even more restrictive in terms of assigning the segment as “real,” but I would make notes because it would help focus where I look.

MG8

As you can see, this gentleman only has 4 X matches, and two of those are quite small, One is near 5cM and one is near 7cM, but none of the 4 is compellingly large – meaning they could be identical by chance. I would make a note for the two larger matches by the names of the people he matches, but I would not assign any of the matches as maternal at this point with given the small segments.  When I make “side” assignments, I want them to be as strong as possible so I don’t have to second guess the assignment later.

I would also look at who else I match on the X on that segment and see if there is anything remarkable about common matches.  In males, if the X match is valid, it HAS to come from mother’s side, so if you see a small segment X match in someone you know is related on your father’s side, that’s an indication that the X match on that segment to that person is IBC or you also share a second ancestor on the maternal side.

Larger X matches are already mathematically triangulated for you, meaning, if you’re a male,  you know what side they come from.  There are no “2 sides” to this chromosome and all you need to add is genealogy.  Women, you still have two sides, because you inherited an X from both your mother and your father, so you are not mathematically triangulated.  Sorry ladies!

If you would like to read more about the X chromosome, inheritance patterns and matching, click here and here. Be sure to utilize the inheritance pattern genealogy charts.

Summary

See how easy and fun this is when you break it down into easy, logical steps.

Even if you can’t mathematically triangulate, if you can find multiple people in the match group that descend from the same known ancestors or ancestral line, you can form smaller groups with these individuals until you have the opportunity to create a larger mathematical match group.

It only takes three people to create a genealogical triangulation group, so long as they aren’t close relatives.  Siblings don’t count, for example, and neither do parents when counting to 3 in a triangulation group.

As long as you can identify who your DNA on a particular segment came from, you really don’t need to fit everyone on your match list into a triangulation group – so if you don’t have access to some accounts, it’s not the end of the world.  You can also use the ICW tool to determine who people match, in general, if not segment specific..

You’re only going to have two ancestral lines, a paternal line and a maternal line that is relevant to any match group – so once you know who those ancestors are, the rest of your matches HAVE to fall into one group or another, or are outliers.  So don’t obsess about not being able to fit everyone into a match group either mathematically or genealogically.  However, if you can find a genealogical connection, by all means do, because one of those matches may well be the person who isolates the DNA to either the male or female of the ancestral couple and allows you to go back another generation or two in time.

Don’t forget about utilizing your Family Matches with assigned paternal and maternal icons.  That’s a great new tool.

It’s fascinating to see which of our ancestors our DNA came from. Finding new cousins by utilizing DNA is exciting as well – and gives us new opportunities to establish family relationships and share research information – opportunities that never existed prior to DNA bringing us together and providing us with that all important introduction.

Have fun.

33 thoughts on “Concepts – Match Groups and Triangulation

  1. Calvin is most likely IBC, but you cannot verify that with the tools at ftdna. He would not show up in your matches’ ICW lists unless he shared other larger segments with them.

  2. Another GREAT post, Roberta! You’ve highlighted the difference between forming groups using Triangulation methods (I call it mechanical Triangulation – no genealogy required), and Triangulated Groups including Common Ancestors. As I’ve studied this over the past 6 years, I’m coming to the conclusion that we may never figure out the true Common Ancestor for every TG. The Ancestor who first formed the whole (mechanical, or mathematical) TG may be pretty far back. “Sticky” segments (actually a relatively small number of crossovers which can subdivide only a very few segments in each chromosome in each generation) allow segments to come down from some distant ancestors. So when we identify Common Ancestors of Matches in a TG, we are usually identifying the Most Recent Common Ancestor (MRCA) we know among selected relatives in the TG – normally close cousins. As such, we must always be prepared to find more distant cousins in the TG with an MRCA who is ancestral to one of the two Ancestors we originally identified with close cousins. You’ve highlighted this expectation very well in your post. Bravo.

    • Thanks Jim. My groups include a combination of every method I can think of to use to give me clues or proof😊

      For some reason trying to make the differences clear was difficult.

      • You did a great job. I struggle with trying to find clear, every-day wording that gets the concepts across. I’m ready to adopt new terms just for genetic genealogy – the existing vocabulary just doesn’t have the right words.
        And although my main process centers around TGs (they have fixed boundaries in each chromosome, and each one represents a specific ancestral line), the end result comes back to genealogy. Genealogy work – and it is work – communicating and sharing. So I also use clustering and the FAN principle to link names to the segments defined by TGs.

      • Have you written about using clustering and the FAN principle? If so, would you post the links and if not, will you talk just a bit about those? Thanks Jim.

  3. I think using the term “two sides to each chromosome” can be confusing. We do have two sides on each chromosome, but both sides of that chromosome comes from one parent. It’s because we have a pair of every chromosome that we get one entire chromosome from each parent. Otherwise I think you have great explanations, and I love that you keep up on all the latest stuff.

  4. Hi- my great aunt is confirmed through her brother’s Y-DNA test 1/128 Native American. On her AncestryDNA results, she gets a trace region of Central Asian. Could this be the Native American?

  5. Thanks for another interesting article Roberta.

    In my master spreadsheet, I have added columns for each of my eight close relatives that have tested. In the cell I type the cM total. Then I can just read across to see if a match shares DNA with them, and how much. I also have a color-coded bar that shows me at a glance which branch of my tree a match is connected to.

    Jim Bartlett had a great piece in Segmentology about mapping chromosomes by generation. I took has advice and I’m now 100% mapped for my father’s parents (only 45% for my mother’s – no half siblings there). I mention it because for the first time I can see fairly easily which pieces are IBC. We all have so many 25cM matches where there are two pieces – say a 19cM and a 6cM – and I can see now just how many of those 6cM pieces are IBC. But I also feel comfortable noting a 3rd cousin’s 4cM segment if it fits the map.

  6. ROBERTA, HELP!!
    You said,
    “At Family Tree DNA, when you download your chromosome matching results, meaning your complete spreadsheet, then sorting into chromosome order (sort by column, end location, start location, then chromosome) your spreadsheet will look like this.”

    I can sort one column (cM largest to smallest), but when I sort the “end location”, the cM column is no longer largest to smallest. I feel like I’m missing a step. Is it possible that my Excel 2013 cannot sort more than one column at a time?

    MG1

    • You only sort one column at a time. CM is not one of them. You sort in reverse order so that the last column you sort is chromosome and the viewing order is chromosome number, start and end which is the opposite of the sorting order. Just do it in the order I said in the article and you will see when you have sorted all three columns one at a time. In one of the articles I gave some excel resources which might help you too.

  7. Hello my name is Linda
    I recently did An ancestry DNA my DNA results show that I am 44% Native American can someone please help me to understand how this works can I be enrolled into a tribe very much interested in finding what tribe my ancestors came from any help is highly appreciated thank you

  8. I think you were reading my mind. At FTDNA, if I do an ICW for a first cousin (maternal), then all of her matches are triangulated to my Mothers side, correct? And taking that a step further, if I do a triangulated match with her at GedMatch, can I combine both groups together in my combined spreadsheet? Also, back to FTDNA, if I do an ICW with all of cousin Karens matches in my match list, does that mean all of those matches are also Mothers side?
    Thanks, Don

      • Sorry, Karen is my first cousin – mother’s side. Have now re-read the sections. Looking for confirmation of my understanding of your dissertation. If an individual is on my match list at a given cH and segment, and that same person is on Karen’s (I have her permission to sign in on her Kit) match list at the same cH and segment, and assuming large enough numbers to be significant, that would put that individual on mother’s side?
        Don

  9. Has the change at FTDNA affected the “in common with” results? There are some people who used to show up as “in common with” and showed overlapping results when entered into DNAGedcom (even though I have no paper trail to one and have a paper trail to another who matches the first with a surname not known to me) but are not now in the “in common with” results. Does that mean that they never were a three-way match or is something else going on?

    Jane

  10. Pingback: Double Match Triangulator (DMT) | DNAeXplained – Genetic Genealogy

  11. Pingback: Ancestry V1 vs V2 – Shared DNA and Relationship Predictions | DNAeXplained – Genetic Genealogy

  12. Pingback: Lazarus Dodson (1795-1861), Under the Radar?, 52 Ancestors #139 | DNAeXplained – Genetic Genealogy

  13. Pingback: 23andMe’s New Ancestry Composition (Ethnicity) Chromosome Segments | DNAeXplained – Genetic Genealogy

  14. Pingback: 2016 Genetic Genealogy Retrospective | DNAeXplained – Genetic Genealogy

  15. Pingback: Concepts – Segment Size, Legitimate and False Matches | DNAeXplained – Genetic Genealogy

  16. Pingback: The Concepts Series | DNAeXplained – Genetic Genealogy

  17. Pingback: Henry Dagord or Dagod or Maybe Doggett (c 1660/1683 – after 1708), 52 Ancestors #150 | DNAeXplained – Genetic Genealogy

  18. Pingback: Concepts – Segment Survival – 3 and 4 Generation Phasing | DNAeXplained – Genetic Genealogy

  19. Pingback: Genetic Communities | DNAeXplained – Genetic Genealogy

  20. Pingback: Which DNA Test is Best? | DNAeXplained – Genetic Genealogy

  21. Pingback: Anne (probably not) Elmore (c1650/2-c1721), Wife of Charles Dodson, 52 Ancestors #159 | DNAeXplained – Genetic Genealogy

  22. Pingback: Concepts – Why Genetic Genealogy and Triangulation? | DNAeXplained – Genetic Genealogy

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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