Tuesday, 21 June 2016

Should I upgrade my Y-DNA test to 67 or 111 markers?

If you have done a Y-DNA-37 test with FamilyTreeDNA, you may be wondering if there is any point in testing a higher number of markers (67 or 111) and what would be the benefit of such testing. Is it worth doing it? And if so, why?

Well the answer is yes, but only under certain circumstances. Outside of these circumstances you might be better spending your hard-earned cash on a different DNA test ... or on your favourite ice-cream.

Here are the main reasons for upgrading to Y-DNA-67 or Y-DNA-111:
  1. No or few matches at 37 markers
  2. Lots of matches at 37 markers
  3. To assist the Project Administrator with difficulties in placing you in a group
  4. To more precisely estimate how closely two specific people are related
  5. To help the PA identify the branching pattern within a genetic group 
We will look at each of them in turn, but before we do let's mention a few key considerations about Y-DNA testing in general, how matches are identified, and some of the pitfalls involved in the process.

Some general considerations

I choose 37 markers as the starting point because most people interested in surname research will have tested to this level. Not everyone will however, and some people (especially transfers from the National Genographic Project) will only have tested to 12 markers. Neither of these are particularly useful for surname research (with rare exceptions) so I will only be addressing upgrading from 37 markers to 67 or 111.

Secondly, it is important to be aware of FTDNA's threshold criteria for declaring a match and listing them in your Matches List. These thresholds are based on Genetic Distance (GD) and are illustrated in the table below (see FTDNA's FAQ page and Privacy Policy page).  Having a GD of 4/37 means that the two individuals being compared are 4 steps away from an exact match (which would usually be expressed as 0/37, or sometimes 37/37).

The thresholds for declaring a match can be summarised as: having a GD at or below 1/12, 2/25, 4/37, 7/67, and 10/111. Each threshold value roughly equates to about 10% of the total number of markers. 

It is important to be aware that some people who fall within these thresholds will not be related to you within "a genealogical timeframe" (which we will take to be about the last 1000 years or so). Similarly, some people who fall outside these thresholds WILL be related to you "within a genealogical timeframe".

Also, it is important to appreciate that these thresholds are arbitrary. They are designed to maximise the number of true positives (high sensitivity) and minimise the number of false positives (high specificity). However, some true positives will escape being caught and some false positives will sneak through. And one or the other scenario may affect some people more than others. The question is: how do you recognise this? How do you separate the wheat from the chaff? Your chances of being able to do this are substantially increased by joining the appropriate surname and haplogroup projects and liaising with the Project Administrators because they have better oversight of the totality of the data within a genetic group and also have additional tools that they can use to better define how closely you are related to other people.

Interpreting Genetic Distance is just as arbitrary as defining a threshold for "declaring a match" and our thinking on this subject is likely to change over time. The table below is derived from FTDNA FAQ pages relating to Genetic Distance at 12, 25, 37, 67, and 111 markers respectively. Match Thresholds are highlighted in yellow.

There is some apparent inconsistency at the 111 marker level when comparing the Match Threshold (>/=10) to the interpretation of Genetic Distance (Not Related). If two people with a GD of 10/111 are Not Related, why declare them as a match?

Furthermore, with the advent of SNP testing and our increasing experience from surname and haplogroup projects, there is now strong evidence that these interpretations can be wildly wrong. Even two same-surname individuals with a GD of >10/37 could be related within a genealogical timeframe (Farrell DNA Project, group R1b-GF2). The interpretations above should therefore be used only as a guide.

So now let's look at the specific scenarios where it might be worthwhile upgrading. What follows expands on the advice already given by FTDNA in its FAQ pages.

Scenario 1:  No or few matches at 37 markers

If you have (say) no matches at the 37 marker level, it could be because someone has a Genetic Distance to you of (say) 5/37 ... in other words, there are 5 differences between you both in the first 37 markers. However the threshold for "declaring a match" is 4/37, and so neither of you will appear in the other's Matches List.

But if you both upgrade to 67 markers, and there are no further differences between you on markers 38 thru 67, then the number of differences remains at 5 and the Genetic Distance is written as 5/67, which is above the threshold for declaring a match and thus you will each appear in the other's Matches List.

In short, upgrading to 67 markers has revealed an additional match that was "hidden" at the 37 marker level.

The same scenario may also apply at the 111 marker level. But the big caveat is you can only compare yourself with other people who have upgraded to at least the same marker level. You cannot detect more matches by upgrading to 67 markers if everyone else is still at 37 markers. Of the 238,000 people with Y-DNA-37 data in the FTDNA database, only 33% of them have Y-DNA-111 data.

There are several reasons for why you may have no or few matches:
  • you may be the first person with your Y-DNA signature to do the test
  • your DNA signature may be very rare because you are the last of your line, or few people with that particular signature are left in the world
  • you may have unusual mutations which have moved you away from the rest of your group

Scenario 2:  Lots of matches at 37 markers 

If you have lots of matches at 37 markers, either your Y-DNA signature is very common in the population or you are a victim of Convergence. This is where, just by chance, people have a similar genetic profile to you that makes them fall within the matching threshold, but the common ancestor is 1000's of years ago rather than 100's of years ago. 

Upgrading to higher marker levels will help weed out many of these Convergent matches but may not eliminate them completely. Convergence has been observed with a GD of 3/111 in the Stewart DNA Project (see this YouTube video from 28:50 onwards).

Scenario 3:  To assist the Project Administrator with difficulties in placing you in a group

Sometimes it can be difficult to allocate project members to a specific genetic group within a surname project, for example if the GD is borderline (e.g. 5/37) and/or the member has a surname variant that may or may not be related (e.g Farrell and Harrell).

In these circumstances upgrading to a higher level of markers may provide additional supportive evidence for grouping you in a specific group (e.g. if the GD remained the same at 67 markers, namely 5/67, then this would be stronger evidence for including you in a specific group).

This scenario may be particularly relevant to you if you are in the Ungrouped category in a surname project. If so, ask your Project Administrator if upgrading to 67 or 111 markers would help him or her with the grouping process. 

Scenario 4:  To more precisely estimate how closely two specific people are related

Upgrading to 67 or 111 markers can help provide supportive data of a very close relationship on the direct male line. However, this should probably be done in conjunction with autosomal DNA testing (and even mtDNA testing) as the Y-DNA-111 test on its own is not conclusive.

FTDNA says that over 50% of exact matches at 111 markers (GD = 0/111) are first cousins. Similarly, over half of matches with a GD of 1/111 are 2nd cousins or closer, 2/111 are 4th cousins or closer, 3/111 are 5th cousins or closer, and so on (see full Table here). 

In short, upgrading to 111 markers will give you a better estimate of how close you are related to someone else but will not define it precisely. There will still be quite a broad range around the "best guess". In order to get a more precise estimate of which ancestor on a direct male line is the common ancestor between two people, it may be necessary to do autosomal DNA testing to estimate the degree of kinship, or to additionally test specific selected cousins of one or both matches, in order to triangulate with atDNA testing, or even mtDNA testing (the latter technique was used to identify WWI soldiers found in Fromelles).

Scenario 5:  To help the PA identify the branching pattern within a genetic group

As surname projects mature, some Project Administrators may take on the task of better defining the branching pattern within certain genetic groups within the project. I am attempting this in the Gleason/Gleeson DNA Project (you can see more about it in this YouTube video).

This process of building a Mutation History Tree (or cladogram or phylogram) is not easy and requires a lot of work. It is best done with 111 STR marker data combined with SNP data (e.g. via the Big Y test). In the future, the number of STRs available to test may increase to 500 or more (e.g. via YFULL) and testing out to 500 markers may become the preferred option. Furthermore, this process requires that many people within a genetic group have this data available.  It is thus quite a costly undertaking for group members.

However, defining the branching pattern within a genetic group brings several specific potential benefits. It can more accurately define how long ago different branches of the family broke away from each other, and how closely specific individuals within a family are related. This can be very useful for both historical studies of the family and the personal genealogical research of individual members. It can also indicate where Back Mutations and Parallel Mutations occurred within a particular genetic group, and this furthers our understanding of the nature of these mutations which usually remain hidden.


So if you think you fall into one of the above categories, consider upgrading your Y-DNA-37 results to the 67 or 111 marker level. You can do it in a step-wise fashion as there is (usually) no extra cost in doing it this way rather than upgrading to the highest level all at once.  And this potentially saves you money because all your questions may be answered by simply upgrading to 67 markers only.

If you do not fall into one of the above categories, you may benefit more from some other test, such as Y-SNP testing or autosomal DNA testing. It all depends on the questions you want answered.

Defining the genealogical questions clearly in your own head will enable you to better arrive at the optimal testing strategy to answer your questions.

Maurice Gleeson
June 2016


  1. Well, done, Maurice.

    1. Another reason for few matches is belonging to an ethnic or geographic group where people have not tested. The "I-know-who-I-am" crowd.

    2. Do you have a probability table for Y-67 tests similar to the one you show for the Y-111?

    1. Good point, Israel. And you can find the Y-67 Table by clicking on the 67 link before the second image in the text above... or here - https://www.familytreedna.com/learn/y-dna-testing/y-str/two-men-share-surname-genetic-distance-67-y-chromosome-str-markers-interpreted/

  2. Maurice, please consider two additional reasons to upgrade to 67 and/or to 111. The first is to identify STR signature markers, which may help guide future testing and place you in the correct cluster. This can save you money on future SNP testing as well as help you identify who you should recruit for additional SNP testing and as potential people to work with in funding additional SNP testing.
    There are many STRs that are critical to cluster groupings that are not in the first 37. My own group, R1b-L513, has a very consistent DYS406s1=11or12 and DYS617=13or14 which are higher than the modal for the overall haplogroup. This is how I started building my first project, finding these people and it has proven out that we have about 800 people identified now.
    Another example is DYS492=13 which is mode for R1b-U106 but is rare in brother R1b-P312 people. If you have DYS492=13, you should check other markers too but you might skip right to deeper testing of SNPs found only in U106.

    1. Excellent example Mike - thank you.

      To clarify, the identification of such "signature markers" for a particular genetic group within a project would only be possible via the Project Administrator - this is not something that the individual would be able to do on his own. This points to the need for close liaison between individual participants and their Project Administrators, particularly haplogroup project admins, who have a much greater overview of the totality of the data than individuals or surname project admins do.

      I would imagine that usually it is the haplogroup project admin that identifies the "signature markers". It would be necessary for that Admin to then notify the Admins of the relevant surname projects impacted by this "signature marker" discovery. But how often does that happen? Is there close enough liaison between surname and haplogroup project admins? Or do they suspect that they do different things and there is no synergy between their various activities? What do you think?

  3. The second reason is the "killer". If one is deeply interested in this hobby you don't want to miss what you could have easily known.
    In all cases, if you don't upgrade beyond 37 STRs, you may miss a match just beyond the FTDNA match limits at 37 that falls in place at 67. Remember, you are almost doubling your marker set when going to 67 so a lot can happen in the 38-67 STR panel. Same goes for 68-111, which is actually a smoother behaving set of STRs (a mini-reason to go to 111).
    We don't know what we don't know which is why we test ... so I ask the question a different way:

    Why wouldn't you go to 67 and 111 STRs?

    You'll never know what you might be missing unless you test. The only reasons not to upgrade are budgetary. Do I have the money and do I care? I submit that many hobbyists spend more on their monthly phone or cable TV bills than one of these upgrades. I shutter at some of the restaurant bills I pick up at family nights out. Anyway, budgetary reasoning is personal so it is up to each of us, but
    if one is passionate about genetic genealogy you might as well bite the bullet. Most of who post have done so.

    1. Good point Mike. If money isn't a major problem, then there really is no reason not to upgrade. It's a hobby, so you are entitled to spend money on it (just as teenage children used to buy CD's or now buy music or Apps for their "devices").

      In addition, upgrading to 67 and 111 markers benefits others in the community ... but only if people have joined the appropriate surname, haplogroup & geographic projects! So I encourage all surname project administrators to identify those haplogroup and geographic projects that might benefit from their project participants membership and then exhort their project members to join them.

  4. Thanks for writing this article which is very timely for me who is confused about the benefits of upgrading when I'm not getting any close matched on the Y37. However I've got an NPE so I'm still skepitcal that I'm investing in a dead horse! I have also done the autosomal to no avail. The sample is 9 years old - should I worry about sustainability? What do you suggest?

    1. Hi Lesley, contact FTDNA about the sample. They may send you a new kit free of charge in order to collect a fresh sample.

  5. I appreciate your article. It seems to me that no one cares about STR testing anymore (except for the financial benefit!) since my Y67 test classes me as M269 and that's it.
    My previous test at 23andme autosomal showed me out to L48, so what's up with that?
    It is all SNPs after that,apparently. I noticed recently that my Stevenson surname group no longer has the STR charts posted, just SNPs that have been tested. Family tree DNA is all about the money,in my opinion. I did 3 different family surnames and have found 2 total close matches that I am still unable to connect on paper!
    IF these project managers are doing anything, I would be surprised.

  6. I appreciate your article. It seems to me that no one cares about STR testing anymore (except for the financial benefit!) since my Y67 test classes me as M269 and that's it.
    My previous test at 23andme autosomal showed me out to L48, so what's up with that?
    It is all SNPs after that,apparently. I noticed recently that my Stevenson surname group no longer has the STR charts posted, just SNPs that have been tested. Family tree DNA is all about the money,in my opinion. I did 3 different family surnames and have found 2 total close matches that I am still unable to connect on paper!
    IF these project managers are doing anything, I would be surprised.

  7. This is very much valuable content, i am a newbie to this and thanks for sharing this info.

  8. It’s a nice article. This is very helpful to me. Thanks for sharing such informative post.

  9. I would like to use genetic testing to learn about my ethnic ancestry. My family is from Guyana and the inhabitants are descendants of indentured servants and slaves brought from Asia and Africa. There are no ancestral records available. I don't particularly care about finding current relatives. Would it be beneficial from me to upgrade to the 67 or 111 markers test? Thanks in advance

  10. Thanks for explaining this so clearly - just what I was looking for. I've got a test that has few matches at 25 and none at 37, and am looking for an adoptee... it seems based on this info going higher will likely not help me because even if I find some matches in 67 I didn't find in 37, they will be VERY distant.

  11. My father has tested to 67 and I am wondering if I should get his 111 ordered or the big y. Im concerned he only has enough sample left for one test or the other, and ordering another saliva sample is not an option. His 6th great grandfather is unknown and I'm trying to pinpoint which person it is.