23andMe’s New and Improved Maternal Haplogroups

It’s springtime for 23andMe’s Maternal Line tree!

Each one of us can trace our ancestry back through our mother, our mother’s mother, her mother, and so on. And there is one piece of our DNA that has been inherited along that exact, maternal line.

By examining mutations in this segment of DNA, called mitochondrial DNA because it resides in parts of the cell called mitochondria, we can learn about our line of maternal ancestors – in particular, where these ancestors came from. Sets of similar, closely related maternal lines are called “haplogroups” and are given labels such as “J1c3” or “X2b”. Numerous researchers around the world study these maternal haplogroups to better understand where they originated and where people with the haplogroups have migrated over the course of thousands of years.  

23andMe interprets the haplogroups for customers based on the publications of these researchers. All maternal haplogroups are themselves related; the relationships are represented as a “maternal line tree”.   23andMe’s version of that tree has just experienced both pruning and new growth via an extensive update.   With these changes, a customer’s particular branch (maternal haplogroup) may or may not have changed, but, overall, this new tree is far bushier, with over 1500 different haplogroup assignments possible, and over 870 haplogroups actually assigned to customers.

Why update the haplogroups?

Maternal haplogroups are of greatest interest if (a) they have a known history or (b) they enable people to discover others who are related to them. Scientists are continually revising the tree that indicates which mutations define each maternal haplogroup, and how all the haplogroups are related. The more up-to-date the tree used for haplogroup assignments is, the more accurate and more meaningful those assignments are. Up-to-date haplogroups enable our customers to answer questions such as the following:

a) Which haplogroup histories are relevant to me? Your maternal haplogroup and the maternal haplogroups of your ancestors provide insights into your ancestry. Let’s say your haplogroup under 23andMe’s previous tree was H2. The known history for haplogroup H2 is very general; women with this haplogroup survived and reproduced very successfully in Europe during the Last Ice Age. Let’s say your haplogroup under 23andMe’s new tree is now H2a2. While H2 is common throughout all of Europe, research has shown that H2a is concentrated near the Caspian Sea. The history for H2a2 has yet to be written, but eventually we’ll understand, possibly through 23andMe’s research, the origin and migration history of this haplogroup that branches off from H2a.

b) To whom am I most similar? Who are your maternal line cousins? In general the higher resolution the haplogroup, the more meaningful the sharing of a haplogroup. For example, if your haplogroup is H2, you share that haplogroup with over 600 23andMe customers; on the other hand, if your assignment is H2a2b, you share that with just over 60 customers. The common ancestor of everyone with H2a2b lived far more recently than the common ancestor of everyone with H2. People with a more specific haplogroup can begin to ask, where are we all from?   Where did our common ancestor live, and when? 23andMe also aims for consistency with other genetic services.   If all such services are using the same haplogroup definitions, then maternal haplogroups are comparable across services.

Steps in the update

In the summer of 2007 when 23andMe developed its initial Maternal Line feature there was no publicly available, up-to-date tree. Therefore 23andMe used information from various sources to create a tree as up-to-date as possible and to assign haplogroups.  

Now there is a publicly available tree maintained by scientist Mannis van Oven of the Netherlands, with help from other geneticists and genetic genealogists (including 23andMe customers!). They update the tree regularly, and we used this tree (Build 7) as a guide for updating our Maternal Line feature. We recalculated every customer’s haplogroup and otherwise modified the feature to be consistent with the tree of Phylotree.org. Many customers will see the same maternal haplogroup as in the past. Thousands, however, will see something new.  

Customer haplogroups have changed for two main reasons: (1) because of new definitions of the haplogroups, and (2) because of improved SNP results. Some customers will see a higher resolution haplogroup.

For example, over 100 customers see a change from H4a to H4a1a. Other customers will see a major change in their haplogroup name, even though the history of that haplogroup has not changed. For example, some customers who had an H5a assignment will now see H36.   Others who had PreHV1 will now see R0. Some customers will see a less dramatic change: all customers with a J1a haplogroup will now see a J1c haplogroup. Many of these changes simply reflect relabeling of the tree by van Oven and other scientists. The most common change is the removal of a “*” from a haplogroup label. The “*” is used to indicate that a person’s haplogroup cannot be assigned to a more specific subhaplogroup (for example H* means that a person is not in H1, H2, etc). As the maternal haplogroup tree grows, the “*” will necessarily change in meaning, so 23andMe has decided to do away with it and instead use a system in which we just assign people to the most specific haplogroup possible.  

So if you are definitely in haplogroup H, but not a part of H1, H2, etc , then you will simply be assigned to haplogroup H. In some cases a haplogroup has been split. So some people with T2b2 will now have T2b4, while others continue to have T2b2. Sometimes customers will have less detailed haplogroups. For example, a few hundred customers go from T2b2 to T2b because T2b2 was redefined.

Send us your feedback!

With over 870 distinct haplogroups assigned to more than 35,000 customers, we anticipate making further refinements.   If you have information regarding your mtDNA that is inconsistent with our update, please let us know at mito-feedback@23andme.com!

Reference: van Oven M, Kayser M. 2009. Updated comprehensive phylogenetic tree of global human mitochondrial DNA variation. Hum Mutat 30(2):E386-E394. http://www.phylotree.org. doi:10.1002/humu.20921

  • Lisa Boyer

    Yay!!! I hope I become asteriskless! Our U5b1b* is the biggest mystery ever. I was supposed to have Cherokee mtDNA…

  • Aussie

    This may give many of us a new challenge to understand who our mum really was. Yay two.

  • Elizabeth (rainbow)

    My 23andme H1* now says H1c. 🙂
    I already knew I was H1c because I have 477c.
    I am waiting for my FGS from FTDNA to find out if I am H1c1 or H1c2 or H1c3.

  • MJ

    LOl–could have save a lot of money if they had done this sooner! The new group agrees with FTDNA FGS testing results.


  • This is a nice update, but why Build 7? Build 8 is the current tree. I was hoping you could keep things updated as each new tree is released rather than once every 3 years.

  • Joanna Mountain

    @James Lick

    We are excited to have updated to Build 7 of Phylotree; Build 8 was released too late to be included in this update. However, we will be looking for ways to update this feature more often.

  • Mike Zapolski

    Unfortunately my maternal haplogroup is presently shown as H36, which I believe is incorrect based on my personal mitochondrial mutations. While I am in the H branch of the ‘tree’, my personal mutations place me in the H5b haplogroup, and prior to this recent revision 23andMe had me assigned to the H5 group. Ooops…

  • Eric Uhden

    I was K1c1b at first and now I am just K1

  • JamyRogers

    I am still a plain ol’ H.

  • Marsie Hofmann

    I also was a K1c1b and now I am just K1, perhaps this is a misprint.

  • William Wale

    I applaud your efforts to improve information about specific haplogroups, though I am still very disappointed to have only very general information about haplogroup H (changed from H*). This tells me little or nothing about my maternal line and even that in very general terms. I am quite disillusioned about this fact, hoping for and expecting much more.

  • sc

    They need to re-classify the paternal side too. They still have mine listed as G2a5 which hasn’t been in use for over two years. Having the correct classification would help when searching for relatives.

  • Wendy

    Elizabeth, I’m H1C too! You sound very knowledgeable. I’ll try to find you on 23andme but i’m wendyallyson.

  • Alexander Inglis

    I am thrilled to have migrated to J1c3. But suddenly I feel all alone … !

  • Richard Upchurch

    I have been updated backwards from H13* to H. In the meantime I just received my FGS from FTDNA and am designated H13. I have also been accepted by GenBank as H13b (a very rare designation).

    FTDNA does not have the H13b designation yet. I am hopeful that my results will help FTDNA add H13b.

    So my question is are you moving me to H temporararily or have you missed something?

    I will be happy to communicate privately with any of your researchers and give you more information about my results . I question whether moving me out of H13 to H, if permanent, is an improvement.

    Please advise,

  • Joanna Mountain

    We at 23andMe appreciate your informing us of either your excitement or your disappointment about the recent update to our Maternal Line feature. We share the excitement of the thousands of individuals who see more detailed haplogroups for themselves or their shares. For those of you who have such a detailed haplogroup that you feel lonely, we wish you luck in finding people with the same haplogroup. For those of you hoping for more detail, we hope that both 23andMe and the research community more broadly can make the discoveries to bring you that detail. And for those who see the need for corrections, rest assured that we are recording the suggestions and acting upon those.

    If you have additional details regarding possible corrections, please send a message to mito-feedback@23andme.com.

  • Mike Zapolski


    Thank you for your note, however it would be a nice gesture if 23andMe could individually confirm the fact that miss-assigned haplogroup emails sent to “mito-feedback” were being investigated and corrected.

    Recognizing that a reply takes some effort, hopefully 23andMe hasn’t received that many problem emails so the courtesy of a reply wouldn’t represent a burden to the staff.

    Nonetheless, it is comforting to know specifically that one’s issue is being addressed…

  • Joanna Mountain

    @Mike Zapolski

    We understand the interest in getting a response to information sent to mito-feedback@23andme.com, and are working to send such responses within 5 business days.


    I am a new 23andMe member but disappointed in the “new and improved” Haplogroup. At FTDNA I am H10 but only H at 23andMe. Hopefully soon they will have a New, New and Improved!

  • kendra goodin

    I am doing a tree for my cousin..he was adopted..
    we had him tested w/ Family Tree..in Texas..
    may I enter his information on Ancestry..
    and then possibly get further test throug Ancestry?
    His mtDNA haplo group is H.. and his Y is R1a1b2.
    We just got the results a week ago.

  • @ kendra goodin

    Unfortunately there is no way for you to import your cousin’s Family Tree information into a 23andMe account.

    We do, however, offer a way for customers of the Ancestry.com DNA service to import their maternal and/or paternal ancestry results into a free 23andMe haplogroup account. (More here: https://www.23andme.com/about/press/20080909a/)

    If you would like to learn more about our ancestry offerings, please visit:

    If your cousin is looking to find biological relatives using DNA, our Relative Finder feature may be of particular interest:

  • Penny Andrews

    question re W1 grouping….it is not shown on Build 7 tree.
    Do we presume it is N+ and then my question goes to
    why isn’t it shown?

  • Carol

    I’m T1a…. and need to ask again why did you did away with the a part ?
    I have one perfect match based on this a. There are many with T* and T1 and T2 that do not match my DNA.

    Carol would like to hear from more T1a matches – only if they have taken the full sequence test.

  • # Alexander Inglison 10 Apr 2010 at 7:08 am
    I am thrilled to have migrated to J1c3. But suddenly I feel all alone … !

    I am also a J1c3 Alexander

  • Irena

    It looks wrong. My hg was changed to W and my daughter stayed at W1.
    We have identical mutation at i4001186 so it doesn’t make sense.
    You did half job. That’s bad.

  • @@ that would rock! i want to be more than just a K. K period. i mean… what does that give me details of… generalization.

  • Richard Upchurch

    My real curiosity is this; there was something about my results that caused 23andme to correctly designate me as H13* when I was first tested.

    Question #1
    What has changed in my results for 23andme to no longer be able to correctly classify me as H13*?

    The answer; Nothing
    My results are the same.

    Question #2
    What has changed in the 23andme classification methodology to cause 23andme to no longer be able to correctly classify me as H13*?

    Please research the answer to #2 and correct this obvious error.


  • Kim

    Is anyone else H6a1???

  • Leanne

    Hey I’m haplogroup T and adopted. I’m desperate for answers. I need to find out what my heritage is and ancestry.com gave me a very broad view can anyone help? please

  • Maggie

    I am H6a1. I’m adopted and trying to find relatives with
    dna. I have one potential 3rd-5th cousin, listed as 4th cousin probability with 65% Shared DNA and 2 shared segments and am not sure how close we really are. I think that 65% with 2 shared segments have got to be real close. When I mentioned adopted she clammed up, so I think she might know that we are close.

  • Fredrich

    Is anyone else H24 I would like to know more about his group.Geographical locations and frequency. My test is pending so I am starting to ask questions. Is there an H24 club?

    • Hi Fredrich,

      Be sure to check out our community forums on haplogroups. (You will need an account set up to view them, which you should have set up when you claimed your kit.) For example, this one may be of interest: https://www.23andme.com/you/community/thread/6281/

  • Audrey Valente

    I am a maternal line DNA halygroup of T2b2 but I can find very few that have the same halygroup.

    Could you explain to me why this is so. I have read that T2b2 has been changed, but can you give me some explanation of background since I can find so little about it.
    Audrey Valente

  • BBB



  • Liz S.

    For Audrey Valente,

    I just found out that FTDNA says that I’m T2b; but through both Pike & Phylotree – I see that I am T2b2.



    I haven’t found anyone with my specific mutations posted on the net! Lol, I’ve been researching as much as I could…there aren’t many of us out there.

    T2b2: is in the cluster 11242G.

    I have 16304T, a back-mutation that T2b2 is “supposed to” have. But I notice that not all T2b2 people have it. It is supposed to mean that at one time in our lineage, we carried the T2b mutation (16304C), but later in the lineage, the 16304C back-mutated to a 16304T (mutated back to an even earlier ancestor than the T2b….to T2 or T). When I first knew I had the 16304T, I just thought it was non-mutated (not back-mutated), so I was surprised that I didn’t end up in a T2a label.

    I also have an extra base pair, so I have 16570, instead of 16,569 (315.1C insertion of extra genetic material…not an insertion that replaces a deletion).

    I also lack a few coding region mutations that other T2b2 people are listed as having.

    I hope this helps.

    Liz S.