One of the questions often asked is why triangulation in genetic genealogy is so important.
Before I answer that, let’s take a look at why genealogists use autosomal DNA for genetic genealogy in the first place.
Why Genetic Genealogy?
Aside from ethnicity testing, genetic genealogists utilize autosomal DNA testing to further their genealogical research or confirm the research they have already performed. Genetic genealogy cannot stand alone on DNA evidence, but must include traditional genealogical research. DNA is simply another tool in the genealogist’s tool box – albeit a critical one.
There are three established primary vendors in this field, Family Tree DNA, Ancestry and 23andMe, plus a few newcomers. All three vendors offer autosomal DNA tests utilized by genetic genealogists in various ways. If you want to learn more about the differences between these vendors’ offerings, please read the article, “Which DNA Test is Best?”
In order to achieve genealogical goals, there are four criteria that need to be met. All are required to achieve triangulation which is the only way to confirm a genealogical ancestral match to a specific ancestor.
- DNA Matching – The tester’s DNA matches that of other testers at the company where they tested, or at GedMatch. All three vendors provide matching information, along with GedMatch, a third-party tool utilized by genetic genealogists.
Family Tree DNA assigns matches to either maternal, paternal or both sides of the tester’s tree based on connecting the DNA of relatives, up through third cousins, who have tested to their appropriate location in the tester’s tree.
In the example above, you can see the individuals linked to my tree include my mother with her Family Finder test, plus her two first cousins, Donald and Cheryl Ferverda who have also tested.
- Ancestor Matching – The testers identify a common ancestor or ancestral line based on their previous work, aka, genealogy and family trees. In the example above, the common ancestors are the parents of the brothers, John and Roscoe Ferverda. Identifying a common ancestor is an easy task with known close relatives, but becomes more challenging the more distant the common ancestor.
Of the vendors, 23andMe does not have a Gedcom upload or ability for testers to display trees and for the vendor to utilize to match surnames, although they can link to external trees. Ancestry provides “tree matching,” shown above, and Ancestry and Family Tree DNA, shown below, both provide surname matching.
- Segment Matching – Utilizing chromosome browsers or downloaded match lists including segment information to identify actual DNA segments that match other testers.
Family Tree DNA’s chromosome browser is shown above.
Each individual tester will have two groups of matches on the same segment, one group from their mother’s side of the tree and one from their father’s side of the tree. Each tester carries DNA inherited from both parents on two different “sides” of each chromosome. You can read more about that in the article, One Chromosome, Two Sides, No Zipper – ICW and the Matrix.
Of the three vendors, Ancestry does not provide segment matching, a chromosome browser, nor any segment information, so testers cannot perform this step at Ancestry.
23andMe does provide this information, but each tester must individually “opt in” to data sharing, and many do not. If testers do not globally “opt in” they must authorize sharing individually for every match, so testers will not be able to see the chromosome segment information for many 23andMe matches. In my case, only about 60% are sharing.
Family Tree DNA provides a chromosome browser, the file download capability with segment information, and everyone authorizes sharing of information when they initially test – so there is no opt-in confusion.
Ancestry and 23andMe raw DNA data files can be transferred to both Family Tree DNA and GedMatch where chromosome browsers and other tools are available. For more information about transferring files, please read Autosomal DNA Transfers – Which Companies Accept Which Tests?
Triangulation – The process used to combine all three of the above steps in order to assign specific segments of the tester’s DNA to specific ancestors, by virtue of:
- The tester’s DNA matching the DNA of other testers on a specific segment.
- Identifying that the individuals who match the tester on that segment also match each other. This is part of the methodology employed to group the testers matches into two groups, the maternal and paternal groupings.
- Identifying which ancestor contributed that segment to all of the people who match the tester and each other on that same segment.
In order for a group of matches to triangulate, they must match each other on the same segment of DNA and they must all share a common ancestor.
Triangulation is part DNA, meaning the inheritance, part technology, meaning the ability to show that all testers in a match group all match each other and on the same segment, and part genealogy, meaning the ability to identify the common ancestor of the group of individuals.
The following chart shows a portion of my match download file on chromosome 5 from Family Tree DNA.
As you can see, these matches all cover significant portions of the same segment on chromosome 5.
Without further investigation, we know that I match all of these people, but we don’t know what that information is telling us about my genealogy. We don’t know who matches each other, and we can’t tell which people are from my mother’s and father’s sides. We also don’t know who the common ancestor is or common ancestors are.
However, looking at the trees of the individuals involved, or contacting them for further information, and/or recognizing known cousins from a specific line all combine to contribute to the identification of our common ancestors.
Below is the same spreadsheet, now greatly enriched after my genealogy work is applied to the DNA matches in two additional columns.
I’ve colored my triangulated groups pink for my mother’s side and blue for my father’s side.
In this case, I also have access to my cousins’ DNA match results, so I can view their matches as well, looking for common matches on my match list.
One of the reasons genealogists always suggest testing older family members and as many cousins as possible is because triangulation becomes much easier with known cousins from particular lines to point the way to the common ancestor. In this case, one cousin, Joe, is from my mother’s side and one, Lou, is from my father’s side.
By looking at my matches’ genealogy, I’ve now been able to assign this particular segment on chromosome 5, on my mother’s side to ancestors Johann Michael Miller and his wife Susanna Berchtol. The same segment, on my father’s side is inherited from Charles Dodson and his wife, Ann, last name unknown.
In order to achieve triangulation, the common ancestor must be determined for the match group. Once triangulation is achieved, descent from the common ancestor is confirmed.
Unless you are dealing with very close known relatives, like the Ferverda first cousins, there is no other way to prove a genetic connection to a specific ancestor.
At Family Tree DNA, I can utilize the chromosome browser and the ICW and matrix tools to determine which of this group matches each other. At 23andMe, I can utilize their shared DNA matching tool. This information can then be recorded in my DNA spreadsheet, as illustrated above.
Triangulation cannot be achieved at Ancestry or utilizing their tools. Ancestry’s DNA Circles provide extended match groups, indicating who matches whom for a particular ancestor shown in a tester’s tree, but do not indicate that the matches are on the same segment. Circles do not guarantee that Circle members are matching on DNA from that ancestor, only that they do match and show a common ancestor in their tree. The third triangulation step of segment matching is missing. Ancestry does not provide segment information in any format, so Ancestry customers who want to triangulate can either retest elsewhere or download their data files to either Family Tree DNA or GedMatch for free.
Before the advent of genetic genealogy, genealogists had to take it on faith that the paper trail was accurate, and that there was no misattributed parentage – either through formal or informal adoption or hanky-panky. That’s not the case anymore.
Today, DNA through triangulation can prove ancestry for groups of people to a common ancestor by identifying segments that have descended from that ancestor and are found in multiple descendants today.
Of course, the next step is to break down those remaining brick walls. For example, what is the birth name of Ann, wife of Charles Dodson, whose surname is unknown? Logically, the DNA descended from a couple, meaning Charles and Ann, contains DNA from both individuals. We don’t know if that segment on chromosome 5 is from Ann, Charles, or parts from both, BUT, if we begin to see a further breakdown to another, unknown family line among the Charles and Ann segments, that might be a clue.
One day, in the future, we’ll be able to identify our unknown family lines through DNA matches and other people’s triangulation. That indeed, is the Holy Grail.
If you’d like to read more specific information about autosomal DNA matching and triangulation, be sure to read the links in the article, above. The following articles may be of interest as well:
- DeMystifying Autosomal DNA Matching
- Autosomal DNA Testing 101 – What Now?
- Autosomal DNA Matching Confidence Spectrum
- Concepts – Segment Size, Legitimate and False Matches
- How Your Autosomal DNA Identifies Your Ancestors
- Concepts – Identical by Descent, State, Population and Chance
- Nine Autosomal Tools at Family Tree DNA
If you think you might come up short, because you have only one known cousin who has tested, well, think again.
Here’s wishing you lots of triangulated matches!!!
I receive a small contribution when you click on some of the links to vendors in my articles. This does NOT increase the price you pay but helps me to keep the lights on and this informational blog free for everyone. Please click on the links in the articles or to the vendors below if you are purchasing products or DNA testing.
Thank you so much.
DNA Purchases and Free Transfers
- Family Tree DNA
- MyHeritage DNA only
- MyHeritage DNA plus Health
- MyHeritage FREE DNA file upload
- 23andMe Ancestry
- 23andMe Ancestry Plus Health
- Legacy Tree Genealogists for genealogy research
Thank you Roberta for this informative article about triangulation. I particularly liked that you shared the spreadsheets, showing your work before and after you add the relatives and common ancestors. The pink and blue highlighting is a great visual tool. Due to inter-family marriages (endogamy), I may need blue and pink stripes on some of my matches. LMAO
As triangulation gains popularity beyond those that are professional genealogists, I hope there’s a grass roots effort to teach this process at local or regional locations. Increased participation, and tools like triangulation are keys to breaking down those brick walls in our lineages. Now, if you only could convince the powers at Ancestry to take down the wall blocking users from access to a chromosome browser.
Hey, I’m a genealogist, not a fairy godmother:) Believe me, if I could, I would!! Actually, since most people don’t understand what they are missing, and it is somewhat complex, I doubt very seriously if they will ever entertain the idea. I think that train left the station long ago.
In all fairness to Ancestry, a chromosome browser and it’s function are indeed somewhat complex, and would probably jam the phones with an abundance of customer service inquiries. Someday, the IT engineers at Ancestry will create a user friendly product with a chromosome browser. KISS (keep it simple stupid). It may even be a featured tool that’s included with an imagined future purchase of GEDmatch.
I’m trying to use the FTDNA chromosome browser and feeling frustrated because it isn’t multicolored like the one you showed, and it doesn’t have a legend to explain the color coding. There are only three colors: navy blue, mustard and grey.
The tutorial doesn’t explain the color coding either. . I surmise that the mustard is matching because I’m comparing to my sister, but I don’t know what the grey means.
Thanks for any Light you can shed on this.
I’ve never seen what you are describing at Family Tree DNA. Grey areas are generally are untested regions. If you compare 5 people, you should see a different color for each person that you’re comparing. If you’re just comparing your sister, you would only see the color for her matching segments to you.
Navy blue is just a background color showing the tested regions. Gray, as Roberta said, is untested regions. The mustard color would then be where you match your sister (if that’s who you’re comparing to). For each additional person you include on the chromosome browser, there will be a different color–shown with their names in the boxes to the left. The FTDNA browser will NOT indicate who matches anyone other than you and will not separate out between mother’s side and father’s side. It just shows exactly where your matches match you. This may be more clear if you select several matches before viewing the browser.
Help a sister out, please! My mother and I are tested at 23andme. I keep a spreadsheet but have encountered some things that make me question what I thought was basic.
Any DNA that I don’t share with her must be from my father, correct? I have multiple people with 3 generations tested telling me that the grandparent and grandchildchild share segments that they themselves (the parent) do not. Unless the grandparent is related to both parents of the child, this isn’t possible, is it?
If my mom has five matches that I do not share with, say on Chrom. 1 30000000-81000000, then those matches come from the opposite parent than what she shared with me? Shouldn’t all of those matches match each other?
Not necessarily. There are also false matches. Read these two articles. https://dna-explained.com/2017/04/06/introducing-the-match-maker-breaker-tool-for-parental-phasing/ and https://dna-explained.com/2017/01/19/concepts-segment-size-legitimate-and-false-matches/
If your Mom had matches that you don’t, it only means you didn’t inherit that DNA.
And no, they won’t all match each other because your Mom has two sides to her chromosome and there are also false matches.
Yes it is getting to be complicated for a lot of people. Part of the problem of course is that different companies offer different types of reports. So, how can we compare between them?
I am surprised that you did not talk about Gedmatch. They produce a triangulation report for all results that are on their site regardless of where the testing was done. Their chromosome browser is especially useful for people, and they are more and more numerous, who have tested only at AncestryDNA. I recently was able to put two people in contact one who had tested at 23 andMe and the other one at ancestryDNA. It turns out that they are second cousins. And, in this case, it turned out to be extremely important to one of them, because this person is adopted.
I am working on another adopted person’s data. One of the problem in this case is to find who is the most recent common ancestor. Although we have triangulation data, what is missing is solid genealogical data. I was all excited when I found a common surname. But when I went up at the line of these people, they lead to two different ancestors who share that surname. So at least one of them is wrong. Will they be thankful when I tell them that what they wrote on their tree is wrong?
So I agree with you that triangulation is most important, but do all genealogists care about the accuracy of their data five generations back?
Thank you, Roberta, for a lucid explanation of Identical by Chance. I have read many others’ attempts to explain it, but your textual and graphic explanation is the best I have seen.
This is a phenomenon governed by “chance.” You illustrated it with segments only 10 base pairs long. As a non-expert in probability, it would seem to me that identical segments running to hundreds of centiMorgans in length would have a nearly infinitesimally small probability of arising by chance during the formation of gametes in two unrelated individuals.
Still, it does happen, so there must be something wrong with my inexpert intuition.
Having had this happened to me several times, I think that the phenomenon is more by “population” than “by chance”. Endogamy increases the chances significantly, so does the founder effect, where the gene pool has been limited, but the number of individuals from the same pool is very large. I belong to such groups: Acadians and French Canadians. The number of genocousins that AncestryDNA finds me now exceeds 30 000 and the number of IBS in there is significant, especially in the last 10 000.
Roberta, how far back in time have you been able to triangulate? My attempts trying to TG with 4th cousins, or further back in time, are very difficult, even starting with the same tree match. It becomes evident to me that many trees are wrong, including my own.
I have two lines confirmed with several individuals who are 10 generations back. When the ancestor is that far back, for me to be confident, I need several people descending from different children and with solid pedigrees so we know there aren’t secondary lines in there as well.
Roberta, Another great post! I love your Holy Grail. Using Triangulated Groups, I have mapped most of my DNA – that is I know where many of the DNA segments from my ancestors start and end. So I know the outline of virtually all of the puzzle pieces in a genealogy time frame. The only genealogy needed was to be able to assign each TG to one side or the other. The part that remains – linking each TG segment back to an Ancestor, is all genealogy work. I now know about 75 percent of my TGs back to at least the grandfather level, and back to 5G grandparents (6th cousin level) on some. As I gain confidence (by confirmations from multiple Matches) that the Common Ancestor is correct, I’m able to help my Match cousins focus on specific lines – and we sometimes find the same Ancestor in their Tree. As more and more folks develop chromosome maps (or even confirmed TGs), we will really be able to help each other with solid genealogy. Thanks, again, for teaching this process.
I was wondering recently how much of your chromosomes you have mapped now. 75% is a lot. I should spend less time writing and more time mapping:) I have two lines that are both confirmed to 10 generations. Both lines are descendants of colonial men who had very large families and each generation did too, so there are a lot of people tested from each line. Of course, one can never be absolutely positive there isn’t another relative in the lineage someplace, but with several people all matching on the same segment, and robust trees, it’s very unlikely, especially given descent from different children.
Actually I have 95% of my DNA “covered” with Triangulated Groups – so any new Match easily fits into an existing TG. I’ve determined Common Ancestors for 75% – of course some are more confirmed than others. Like you, I can do a happy dance when several spread out cousins confirm the same CA line. As the Matches continue to pour in (at least doubling our list each year), the TGs will swell with closer and distant cousins. The bottom line is the odds are increasing every day that we’ll find more and more cousins that help confirm our chromosome map.
I read your recent article about the survival of segments of different sizes across three generations with interest. Now, looking at triangulations again – I am wondering if there is any industry standard for what constitutes proof? How many people need to be triangulated? How closely related should (or shouldn’t) they be? Also, is there a minimum size for the segments to be accepted as proof? I have based a few holes in a couple of brick walls, but I’m not sure yet that I’ve totally toppled any. In my case, one wall is 5 generations back from me. “Conventional wisdom” has held for many years that his parents were Randolph Boling and Jane Graves. However, I have never found any paper documentation for that. I have multiple DNA matches to descendants of other children of Randolph and Jane, as well as some matches to descendants of Randolph and Jane’s siblings. I am now working on getting segment data (several matches are at ancestry) and triangulating segments. But when do I have proof? Thanks!
Those are all really good questions, and no, there are no standard answers. If you asked different people, you would get different answers from all of them. In my mind, it’s a combination of segment size, number of people matching and the pedigree chart. With larger segments, I’m much more comfortable with fewer people in the triangulation group – especially with strong pedigrees meaning no unknown people. Of course, we would all agree that close relatives, meaning siblings, parents, aunts, uncles and grandparents can’t be the third leg. I do use first cousins, but I’m much more comfortable if there are multiple people so they aren’t a “foundation” leg.
Triangulation has also helped me to convince skittish matches to share info about their trees. Multiple matches to an email managing accounts sometimes doesnt seem to be enough.
I was also skittish/non believer in Native American ancestry and was convinced by triangulation. I spoke with a DNA match who is a professional genealogist and she immediately triangulated my DNA with other kits she managed. She saw that we all triangulated on a known and proven chromosome segment that is Native American. Until then I was skeptical and thought my small ethnicity percentage was just noise.
Hopefully that response from Bob Harris about Ancestry purchasing Gedmatch is not based on any insider info and is nightmare speculation on his part. I cant imagine anything worse for researchers and not sure why a genealogist would want to see it. Ancestry is currently the most expensive DNA testing service to use all of their tools and the added tools from Gedmatch would most surely not be free and probably hike the price of their monthly subscription.
Personally I would love to have the ability to use gedmatch’s tools automatically on all my ancestry matches. It’s sometimes difficult to even get a response from matches, let alone get them to upload data to gedmatch! However, Benjamin is exactly right in what he says about cost. NOTHING seems to be free at ancestry! I’m pretty sure the additional charges would put those tools out of reach for many people, including myself.
My concern would be that they would do what they did to the Sorenson data base, and their own Y and mtDNA data base – kill them.
Bob Harris, I agree that if Ancestry had a chromo browser, it would jam the phone lines. They have done 4 million dna tests. That is a lot of people. The learning curve trying to grasp and understand the science of dna is STEEP! It would be an impossible nightmare to educate the phone staff at Ancestry, who would then have to educate the public. Impossible to have a phone staff that large. Common sense says Ancestry will never have a chromo browser. Why would they shoot themselves in the foot……JMHO
Enjoyed this post, and I recommended it to my DNA Interest Group as a nice complement to a talk I gave yesterday. Good information for us.
I may have misunderstood you in your article, but you said 23andme users cannot upload their raw DNA files to GEDmatch, however, I just did this last week and had absolutely no problem whatsoever. My 23andme test kit is V3 (pre-2013).
Not sure what you thought you read, but here’s a quote from the article. “Ancestry and 23andMe raw DNA data files can be transferred to both Family Tree DNA and GedMatch where chromosome browsers and other tools are available.”
Yes, I saw that after re-reading. What confused me the first time was the quote, “Of the vendors, 23andMe does not have a Gedcom upload…” so I apologize for the misunderstanding.
Gedcom and GedMatch are often misread for each other. I do it too.
Thank you for your very helpful articles! I could really use some guidance on what to do for my situation:
My father had low/weak sperm in the late 1980s so both my 2-years-older-than-me brother and I were conceived with (different) donor sperm and IVF. My mother’s doctor told her they “mixed-in” some of my dad’s weak sperm with the donor sperm. I also have another older brother who is my dad’s biological son.
My parents know nothing about my sperm donor, but we are pretty sure my dad isn’t my biological father (we are ordering a paternity test to find this out, but still don’t know what kind yet).
I tested on 23andme back when their test was more comprehensive (V3) and recently uploaded to GEDmatch. My mother has agreed to do testing now so we can compare our DNA which may help to identify which DNA of mine is maternal / paternal, but we can’t decide which test she should do.
Unfortunately, we don’t have a lot of money to spend on testing. (I’d like to try to triangulate by having both brothers and my dad test, too!)
Can you recommend what test my mother should do? Should she do FamilyTree DNA and upload to GEDmatch for us to compare that way? Or should she do 23andme just to make it easier to compare with me even though it’s a different version test now? Or should we really bite the bullet and both do FamilyTree DNA?
Thank you so much once again!
The question is your goal. If your goal it to identify your bio father, then you’ll want to test at all 3 because not everyone uploads to GedMatch or anyplace other than where they tested. The FTDNA chip is the same as the 23andMe V3 chip, so that would actually give you better comparative data to each other than the 23andMe V3 vs V4 chip. If your Dad simply takes this test too, you don’t need a paternity test unless you are dealing with a legal situation.
That is very helpful indeed! Thank so much, Roberta. We will be ordering our FTDNA tests today!
Pingback: Concepts – Who To Test for Your Father’s DNA | DNAeXplained – Genetic Genealogy
Pingback: Using Spousal Surnames and DNA to Unravel Male Lines | DNAeXplained – Genetic Genealogy
Charles Dodson and Ann no-name, huh. You’re related to my great-unclt Johnny Dotson. 🙂
Gotta love the Southern Appalachians!
Pingback: Durham DNA – 10 Things I Learned Despite No Y DNA Matches, 52 Ancestors #167 | DNAeXplained – Genetic Genealogy
Can you provide some info about the company/product called MyHeritage? How does it compare to what appears to be your favorite, FTDNA? Thank you!
Yes, I have written about MyHeritage. Please enter the word MyHeritage into the search box at the top right area of the blog and you can then see the articles that include MyHeritage.
Pingback: Introducing The Triangulator | DNAeXplained – Genetic Genealogy
Pingback: 23andMe Automatically Creates Tree Using New Technology – Relationship Triangulation | DNAeXplained – Genetic Genealogy
Pingback: Triangulation in Action at DNAPainter | DNAeXplained – Genetic Genealogy
Pingback: Chromosomes and Genealogy | DNAeXplained – Genetic Genealogy