I recently stumbled across a 2022 FamilyTreeDNA blog post offering suggestions to improve your Y-DNA age estimates for the time to most recent common ancestor (TMRCA). I was intrigued and decided to give the recommendations a try.
Some of their suggestions produced a more immediate change, while others were a bit more difficult to discern their impact. Some were super easy to do while others were a bit more painful to implement. Let me share my experiences so you can determine whether it’s worth your time and effort. (Hint: the quick answer is yes!)
FamilyTreeDNA’s 2022 post made three recommendations to improve your age estimates and help their research team validate and calibrate the paternal Y-DNA tree:
- Specify the birth year for the Y-DNA tester;
- Upload the tester’s known family tree with accurate names and birth years; and
- Link the tester’s Y-DNA matches to confirmed most recent common ancestors.
Birth Year and Family Tree Info
The first two items pretty much go together. Age estimates, or TMRCA calculations, are based on mutation rates for the tested Y-DNA markers (see my earlier post for a more detailed explanation). Because mutations can occur with each new generation, providing your paternal family tree and indicating when the tester was born permits the FamilyTreeDNA research team to use known facts rather than averages and approximations. Y-DNA results become more precise with actual data.
Linking Y-DNA Matches to Your Tree
Providing you know exactly how your Y-DNA matches fit into your tree (and not all of us do), attaching them to individuals in your family tree does two things. First, it reminds you how these individuals are related to you. Second, and perhaps more importantly, it informs FamilyTreeDNA with the same information. This helps them calibrate their estimates for the TMRCA calculations making your branch more accurate and perhaps other testers’ branches more accurate, too. Again, Y-DNA results become more precise because FamilyTreeDNA exactly knows how many mutations and generations separate you and your match from the common ancestor.
Implementing FamilyTreeDNA’s Suggestions
I like experiments, and I’m methodical about it, which is related to my training in the scientific method as an academic. So, I decided to implement FamilyTreeDNA’s suggestions in phases over the course of several weeks so I can see how each suggestion impacts Y-DNA results.
Influencing the parameters of my experiment is how frequently FamilyTreeDNA updates Y-DNA test-related information. They do so once a week, but it can take a couple of weeks for the changes to be visible to you depending on their workload.[1] The changes are observable on their Discover™ tool, which is where we see the Time Tree (where we fit within the larger paternal family tree) and the Scientific Details (the age estimates or TMRCA). Other than our matches and the Block Tree, I believe these items are the most important pieces of information for our genealogical research.
I manage multiple Big Y-700 kits for several of my cousins across my Wilson and Boyd lines. While these test results completed some time ago, for some reason I never added the birth date and paternal family tree for my testers or linked matches to their respective trees. However, this oversight created a good baseline from which to observe the impact in making these changes. During the experiment, I captured images and documented key pieces of information to make the before-and-after comparisons more evident. No new matches posted during the experimental time frame, which limited any potential confounding effects.
Results: Adding the Test Taker’s Birth Year
For my Wilson Y-DNA project, I manage five kits who share a common ancestor from the 1700s. After entering the birth years for the respective test takers in my group, several of the kits shifted on the Time Tree (see the I-Y106972 haplogroup in the image below). Testers are represented on the Time Tree with country flags based on the reported birth location for their earliest known ancestor.
Most of the other test takers’ birth years shown on the above Time Tree are around the year 1950, which is the default value if they do not provide this information.[2] You can see that the birth year for the test takers for three of my five kits is close to 1950, which is accurate as they were born round the 1950s give or take a few years. I additionally have one kit whose test taker was born in the 1960s and the other in the 1920s, which are the ones whose flag on the Time Tree has shifted.
Why is this important? Visually, it gives you an accurate assessment of your matches with respect to your paternal family tree. It may not seem all that impactful now but imagine what it will look like over time as more people test with different birth years. Also, visualizing the generational differences between test takers helps to interpret the number of mutations between test takers as mutations are only possible with each new generation. Mutations are important for TMRCA calculations.
To add a birth year for a Y-DNA tester, login into the account and navigate to the Account Settings in the upper righthand corner of the page. Under the Account Information tab at the bottom of the Contact Information sub-tab is a place to enter the date and place of birth. Also, if you wish for your flag shown in the Time Tree to properly reflect your earliest known ancestor, then go to the Account Settings, then Genealogy, and finally Earliest Known Ancestors. Enter the confirmed country of origin for your earliest known ancestor.
Results: Adding a Paternal Family Tree
Two weeks after I added the birth years, I uploaded complete family trees for the kits I manage ensuring that each direct paternal ancestor had at least a birth year. I did this for my Wilson and Boyd projects. In viewing the Scientific Details within the Discover™ tool, I observed some changes for the TMRCA estimates between my test takers and their common ancestral haplogroup. The table below highlights the before- and after-effects of uploading a family tree.
For illustrative purposes, the Scientific Details for Wilson I-Y106972 haplogroup are shown below for the before (top) and after (bottom) effects for uploading the tree.
Why is this important? While I can’t be entirely certain, I suspect the reason why the TMRCA age estimates for the Wilson I-Y106972 haplogroup adjusted far more greatly than the others is a function of two factors. The first is the number of test takers in the haplogroup. Compared to the other groups, there were more testers, and these testers had greater variation in mutation rates and generational differences. So, the magnitude of possible change was greater. The second factor is related to the first in that without a family tree, the FamilyTreeDNA research team must estimate the number of generations between the test taker and the likely common ancestor. Several of my test takers have long generations, which means the father’s age at the birth of his first child was much older than the average meaning there were actually fewer generations than an approximation would predict. With each tester having varying mutation rates, having an accurate count in the number of generations ensures greater accuracy in TMRCA estimates.
To upload a family tree, login into your account and select the Family Tree tab at the top of the page. Then follow the prompts to upload a GEDCOM file. If you don’t have a family tree from which to generate a GEDCOM file, a family tree can be constructed manually. If possible, it’s best to upload the entire family tree as it will enable you to easily link Y-DNA matches within your tree later, which is discussed next.
Results: Linking Matches to the Family Tree
Interestingly, linking Y-DNA matches to individuals in my family tree didn’t change the Time Tree or Scientific Details. This outcome may be unique to my projects, or it may be that this information may not immediately influence the Time Tree or Scientific Detail charts.
Why is this important? With no immediate change to my Y-DNA results, it might seem this action is not important. However, the reality is that linking matches to your family tree is a long-term strategy. It helps the FamilyTreeDNA research team calibrate their calculations and continue improving the product offering. Just like we’ve become used to our ethnicity percentages to fluctuate overtime due to better data, I believe this is what we’ll experience here if more of us link known matches to our trees.
To link Y-DNA matches to your tree, I recommend two actions: review FamilyTreeDNA’s instructions and be patient. I know this does little to sell the benefit, but the process of linking matches to my family tree was a painful and tedious experience. However, FamilyTreeDNA informs me they are working on a revision to their family tree user interface that promises to make it easier to use, but no time frame for its release has been made.
To begin, go to the Family Tree tab at the top of the home page. Once you see your tree, available matches to link are in the upper left-hand corner separated by the type of test. To link them to members of your tree, use the search bar in the upper right-hand corner to find the relevant person in your tree. This is where it becomes painful. Because the way in which FamilyTreeDNA imports a GEDCOM, the name fields are not as you would expect. Last names are not always searchable despite being visible in your tree. For example, when I searched my tree for “William B Wilson”, it could not see the surname field. Instead it returned anyone named “William”, anyone who has “William” in a birth or death location, and anyone who has a B as a middle initial – all 2,311 of them in my tree. Making the process more difficult, if you had entered a birth or death year as “abt 1809”, FamilyTreeDNA imports this data from your GEDCOM as “???” because it doesn’t know what to do with it (it apparently recognizes “abt.” with the period). The search function is also case sensitive for reasons I do not understand. I don’t think it takes much to imagine my frustration, which led me to contact customer service, who, by the way, were extremely helpful.
An alternative way to search is by birth or death location providing these items are in your uploaded GEDCOM. My William B Wilson was born in Richland, Bucks County, Pennsylvania, and so searching for “Richland” produced only 39 results, which was much easier to navigate in the clunky search process.
Until they release the new family tree tool, I can offer this search hack. If you haven’t yet uploaded your family tree and you know who your matches are that you wish to link, then change the birth or death location for the person of interest in your family tree to a word that appears nowhere else in your tree file. I used “xrick”, which is not a word but is easily remembered and easy to type. I then searched for this “word” to locate the relevant person in which to link the Y-DNA match. Later, I went back into my tree (in both the original GEDCOM and on FamilyTreeDNA) and replaced “xrick” with the person’s correct birth location information. Ugh, the things we do to break through brick walls.
Concluding Comments
I am a huge fan of Y-DNA as evidenced by the methodological rigor associated with this experiment and my other recent blog posts on Y-DNA (“Y” Gift DNA and Big Y-DNA Experiment). I’ve broken down many brick walls and punched major holes in other walls using this outstanding genetic tool. I hope you will give FamilyTreeDNA’s suggestions a try because the more of us that do, the better the paternal family Y-tree will become. And, the stronger the Y-tree trunk becomes, the better will be our respective branches.
Sources
[1] FamilyTreeDNA Blog (2022, September 15), Time Tree: A New Discover Feature. Accessed 7 March 2024 at https://blog.familytreedna.com/time-tree-a-new-discover-feature/.
[2] IBID.