Building Your Personal Mitochondrial Tree

People who test at Family Tree DNA and receive mitochondrial DNA full sequence results often have questions about how they can use their results to further their understanding of their ancestors.

One of the things you can do is to build a mitochondrial DNA haplotree of your own, showing how various people that you match are or are not descended from common ancestors. To do this, you’ll need to contact your matches and share your mutations.

Your results at Family Tree DNA tell you how many mutations you have, shown below, in the genetic distance column.  For more information on genetic distance, how it is calculated and what it means, click here.

GD my results

Your results at MitoSearch, if you upload, or within projects at Family Tree DNA, show you the HVR1+HVR2 region mutations, but the only way to compare the coding region, or full sequence matches is for the people involved to share them directly with each other.

How can mutations help identify your common ancestors with your matches, or if not the ancestor themselves, at least where they were from?

Let’s look at reconstructing a DNA tree based on both your common mutations and mutations you don’t share with your matches.

When building a DNA tree, remember that once a mutation enters the mitochondrial DNA, unless there is a back-mutation, which is exceedingly rare, that mutation will be found in all descendants.

This discussion excludes heteroplasmic mutations, which can be easily identified as any mutation that ends with any letter other than T, A, C or G – for example 16519Y would be heteroplasmic, indicated by the Y. The simple explanation for heteroplamic mutations is that they are a mutation in progress, and therefore relatively recent. They don’t pertain to deeper ancestry, so we are ignoring them for this discussion. Most people don’t have heteroplasmic mutations.

Building Your Tree

Let’s look at an example of how to build a mitochondrial mutation tree.

A common ancestor, at the top of the tree, has 2 mutations that they pass to all of their descendants.

Ancestor B and C have those 2 mutations, so they match ancestor A and each other.

Both ancestor B and C have both developed mutations that don’t match each other. In real life, it would be very rare for mitochondrial DNA to develop mutations in every generation, so just view this as a rather time-compressed example.

In ancestor B’s line, there are two contemporary individuals, D and E, who have all 3 of the mutations that Ancestor B carried.

So, you have a tree that looks like this.  You can click to enlarge.


Ancestor C also has two descendants, F and G, who both carry all of Ancestor C’s mutations, plus both F and G each have a mutation that doesn’t match each other.

So, now let’s say Person I comes along as a match. You can tell which line they belong to, and which lines they don’t, by which mutation(s) person I carries, as compared to your tree. For example, if person I carries mutations 1, 2 and 4, then you know that they are a descendant of Ancestor C, not B.  If they carry 1, 2, 4 and 5, then they descend from Person G’s line.

I suggest that you work with your full sequence matches to build this type of mitochondrial descendancy tree. You must work with your matches, because you cannot see your matches’ coding region results, not even in projects, so you’ll have to ask each one to share with you. Be prepared, some people won’t answer, but often, based on who the people match that do respond to you, and are willing to share, you can figure out the missing blanks.

For example, Let’s say John matches you with one mutation, and so does Joe, but Joe doesn’t answer your e-mail. However, John wants to work with you and John matches Joe exactly. Now you know which mutation Joe has as well – the same one as John.

You know that each of your full sequence matches is within a maximum of 3 mutations difference from you, because that’s the maximum that Family Tree DNA allows to be considered a match at the full sequence level.

Of course, not all of your matches will have the same 3 mutations, which is why you’ll need to work with them to see how your tree fleshes out. Who knows what surprises you may find.

The first question I ask each of my matches, after explaining what I’m trying to do, is whether they share any of my extra or missing mutations, with the exception of the insertions at 309, 315 or 522 and/or any mutation at 16519. These mutations are extremely common. Sometimes people are more comfortable sharing specific mutations than sending you their results. Other people will be glad to send results. In rare instances, the coding region may hold mutations that have medical significance, which is why Family Tree DNA doesn’t show specific mutations, only whether you match or not.


In the example above, you can see that C16189T is normally present in this mitochondrial sequence, but it missing from this person’s results.

Your mitochondrial tree that you build may well shed light on your common ancestor and based on the location of the oldest ancestor of the person at the top of your tree, may also shed light on the location where your common ancestor may have lived and the migration path she took to where your most distant ancestor in this line was found.

My own mitochondrial DNA tree begins in Scandinavia and only my line winds up in Germany before 1700.  Another branch is found in Poland.


Ironically, my exact matches are in Norway (red), not to the line in Poland (orange). The rest of the lines whom I match and that also descend from my Scandinavian ancestor are still found in Scandinavia with one exception found in southern Russia which could be a result of migration to this region from the Germanic region of Europe in the 1700s and 1800s. This tells me that I’m closer, genetically, to the Scandinavian branches than the Polish branch, which is not at all what I would have expected. The Polish branch apparently migrated separately from mine.

My mitochondrial tree also tells me that the common ancestor of all of the matches likely originated in Scandinavia, possibly Norway, also not something I would have expected, given that my most distant ancestor is very clearly German, based on church records.

Give building your mitochondrial tree a try and see what kinds of surprises it may hold!  If you haven’t yet tested your full sequence mitochondrial DNA, order that test today.  You have ancestors waiting for you!



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

Genealogy Services

Genealogy Research


26 thoughts on “Building Your Personal Mitochondrial Tree

  1. So neither the mutation which end with R nor the 309.2C count? Therefore, there really is no tree to build, all my matches are equally likely to be the closest match…

    Not sure if it’s a blessing or a curse…

    • The 309 doesn’t count, but the one that ends with an R might count. It’s considered a mutation in process, so I would ask to see if any of your matches have that one. It’s a good place to start, but don’t be surprised if the answer is no. I would ask anyway because we don’t know what we don’t know.

      • Only one of the matches has it, and we haven’t triangulated her last known ancestress yet, so it could be either insignificant or meaningful, too soon to call.

        Plus we still have three matches who’s ancestress are too recent to be useful which are from the same geographical area that the R mutation match, so maybe they connect, which make the R mutation meaningless, or they are new branches, leaving the R mutation’s case unresolved.

        There’s no way around it, I need to work on it more.

  2. Very interesting idea!

    I have had full sequence mtDNA tests done on the mt descendants of two women with the same maiden surname in the same place at the same time, hoping to establish that they were sisters and eventually find out who their mother was. The results came in T2b5a for both, with a genetic distance of 1. The mt descendant (B) of one of the women has no exact matches. All the exact matches of the other descendant (K) are at a genetic distance of 1 to him.

    Results for tester K
    Extra Mutations: 315.1C 522.1A 522.2C 573.1C T8504C
    Missing Mutations:

    Results for tester B
    Extra Mutations: 315.1C 522.1A 522.2C 573.1C T8504C C16296Y
    Missing Mutations: C16296T

    Is it fair to come to the conclusion that the two women were most likely sisters? (Another possibility, that their mothers were sisters and both married men with the same surname, seems unlikely, but stranger things have happened.) If I understand correctly, B’s extra mutation ending in Y is a mutation in progress and therefore not relevant. The only difference, then, would be that B is missing one mutation and K isn’t.

    Thanks for your insight,

    • 16296 is the same location. What you can say is that it looks like they shared a common ancestor. Do their descendants also match autosomally? In your situation the two tests together could be very powerful evidence.

      • Both testers are the same generational distance from the mystery women (3rd great grandchildren, so would be 4th great grandchildren of the proposed common ancestress and 5th cousins to each other.) B has not taken the autosomal test, but his first cousin (who is not an mt descendant) has and does not match K. However, he does have a few matches in common with K, who are known to descend from K’s mt ancestress.

        For the record, although she is my 2nd cousin once removed, K doesn’t match me autosomally, but she does match both of my sisters (one correctly as a predicted 2-4th cousin and the other as a 4th – remote cousin– so for whatever reason we don’t share a lot of DNA)

        Now I’m going to stick my head out without really knowing what I’m talking about, but I have heard that segments of DNA are more likely to fragment from one generation to the next when the transmission is through females, since crossovers occur at twice the rate in females than males and thus larger block matches tend to be through mostly male lines. If this makes sense to you… could this be a reason why a connection through all females is less likely to show up as a match? I admit to being a bit hazy on the concept, so be kind – my degree is in linguistics, not biology 🙂

  3. Hi Roberta, I have done the basic HVR1 + 2 and have all of 3 matches (Haplogroup U3a) Is it worth doing the full sequence? Will I get more matches that way or is it not worth it?

    • You may get more matches because you must be an exact match at the HVR1 & 2 level but you can have mutations at the full sequence level and show as a match. So if your mutation(s) fall in the HVR1 or two region they won’t show as matches at that level. Many people have more full sequence matches than HVR1&2 but the only way to know is to test. You will receive an expanded haplogroup assignment.

      • Except in my case, I remained the same vague, ambiguous “V” I was at the HVR1&2 level.

        My haplogroup wasn’t expanded, although the details of the assignment were, go figure! Sometimes the past stays maddenly out of reach.

  4. Hi Roberta, these are great ideas for pulling more value out of FMS results Our mtDNA haplogroup project is fortunate to have 48 exact matches but we haven’t yet mapped the mutations of other matches with GD=1 or 2. We will now!

    We can ‘second’ your observation on contacting matches – out of 71 emails sent we were thrilled to receive 44 responses. 22 of them had paper trails to Europe and in each case the oldest known ancestor was from Scotland or Northern Ireland. Sometimes we get lucky.

    One other observation – some matches may exist at 23andMe as they get an estimated haplogroup from the autosomal tests over there. The estimate is often within a few mutations of the definitive answer from FTDNA. I believe those testers can use James Lick’s tool to get a very good shot at their complete mitochondrial tree. They may be be another source of valuable matches with family trees and more folks for these mitochondrial trees. We have found several at mitosearch and on the 23andMe discussion forums.

    Thanks again for this technique. And, by the way, gorgeous quilts at the FTDNA conference!!

  5. Pingback: Family Tree DNA Holiday Coupons and Mitochondrial DNA, The Forgotten Test | DNAeXplained – Genetic Genealogy

  6. Pingback: Mitochondrial DNA – Your Mom’s Story | DNAeXplained – Genetic Genealogy

Leave a Reply