Considering the Odds

 

I've been considering all the encouraging stories about using the tool dubbed What Are The Odds, or WATO. A collaboration between Leah Larkin of The DNA Geek, mathematician Andrew Millard, and Jonny Perl, creator of the DNA Painter website, the tool has generated so many fascinating case studies that I was entranced with its usefulness. The only problem was: I missed one vital detail.

Most of the stories involving use of this tool were applications considering an unidentified parent or grandparent of the person who took the autosomal DNA test. Some were adoptees wishing to discover the true identity of a birth parent. Others were avocational genealogists on a quest to help a parent or other relative from a previous generation find out who their parent might have been. Somehow I lost sight of a fact I already knew: autosomal DNA tests can only reveal so much. The farther back you wander through the generations, the weaker that DNA signal becomes.

In my case, I was working on my husband's DNA test, which was collecting a growing number of matches with people claiming to be descended from a couple named Dennis and Margaret Tully. That seemed fair enough—after all, my father-in-law's great-grandparents had the same names. Better yet, both this couple and my father-in-law's great-grandparents moved from Ireland to what was then called Canada West—present-day Ontario.

That, however, was where the similarities ended. When I discovered the discrepancies—details like dates of birth separated by nearly thirty years, or wife's maiden name not remotely similar—I wanted to find  another way to determine the relationship between what had turned out to be two separate couples. WATO seemed a possible key to testing out some hypotheses.

Scholar-at-heart that I am, I first had to over-study the issue. I read every article and FAQ sheet I could find, then watched webinars explaining how to use WATO, like Jonny Perl's own presentation at Legacy Webinars. I was already ready to launch with a target person and several hypotheses to test, but there was one thing I was missing: my "target person" was not a DNA match; it was the younger Dennis Tully, himself.

Unpacking the question headers on the FAQ page at DNA Painter, I ran across one key detail:

The target person in WATO would normally be someone whose results you have access to.

Because it's necessary to enter in the WATO tool the amount of DNA—think centiMorgans here—the target person shares with each of the others in the family tree, that was my dead stop. There was simply no way I could reverse the clocks, time travel back to the mid-1800s and take a DNA sample from either the senior or junior Dennis Tully. Thus endeth my grand experiment with using WATO to determine whether this centuries-old connection was a father-son relationship, an uncle-nephew one, or something more distant.

Before I simply cave and add the younger Dennis to my father-in-law's family tree in a fit of resigned presumption, I did think of one other way to play with the numbers. Years ago, I had taken a DNA class from Blaine Bettinger—creator of another useful tool, the Shared centiMorgan Project—who had required class members to come prepared for class with the results of DNA tests for three siblings. Because I do not have two full siblings—some of them are half-siblings—I asked my husband and his two sisters if they would provide the test results for me to use.

Once I had those results—and served as admin for the two siblings' tests—I noticed something unusual about the matches each sibling received. This has become a mainstay of my reasons to urge researchers to consider collateral lines. Reminded of the anomaly that showed up in this case, thanks to that original class assignment, I went back to revisit the results of one sister, in view of this latest question about the two Tully ancestors. Since it involves an explanation and some details, I'll save that story for tomorrow's post.