Sunday, July 31, 2022

"What If" Versus "What if NOT"


This week, while muddling over the issue of the Tully DNA match we've been discussing, something else occurred: an unexpected close match showed up on my own readout. Of course, those of us who have tested our own DNA for genealogical purposes need to understand that such a scenario can present itself at any time. When it does happen, however, it can present us with many questions.

This instance involves an adoptee from a previous generation. For a potential parent of the adoptee, I was able to isolate one particular branch of my family, which pointed to three siblings. All three of those siblings are no longer living, so the option of being able to talk to any of them is eliminated.

On the bright side, each of those siblings has had a descendant purchase a DNA test kit at the same company as the adoptee's descendant. Knowing the connection between each of the test-taking descendants and those three siblings, I could look at the amount of centiMorgans shared between each individual family member and this mystery DNA match.

The problem was that none of the test scenarios came with numbers which could rule out any of the siblings. There were other options for solving the puzzle, but they all involved either an outlay of more money or the (for some) terrifying step of tackling the tech approach of uploading data to another testing company to look for other matches there. With that, I was stuck.

The next morning, though, I woke up with a thought. We are so focused on checking out "What If" scenarios—which I did when working with this surprise DNA match—that we forget to also try the opposite. I sat down with the Shared cM Project beta posted at DNA Painter and played with the numbers—only this time, rather than assume a specific sibling as parent and then checking the match with that same sibling's descendant, I played a "What If Not" game.

This approach essentially moved the Most Recent Common Ancestor back a generation. In other words, instead of comparing the descendant of Sibling One with my mystery match while assuming Sibling One was also the match's ancestor, I switched to instead assume the mystery match's ancestor was actually Sibling One's sibling.

Keeping a close eye on the "Relationship Probabilities" chart included above the Shared cM Project readout at DNA Painter was also important. While we all have found DNA matches who push the margins of connectivity, it was helpful to know which possible scenarios produced the least likely result.

Because the Shared cM Project clearly demonstrates there is a range of numbers for each potential relationship, my thought was that—perhaps—I could force one "What If Not" assumption into the gap between two relationships. Perhaps, if the relationships were closer, I could have found that sweet spot.

All was not lost with that experiment. I could see how it cross-applies to the Tully research question I've been puzzling over this past month. For whatever reason, I had strong doubts that my husband's DNA match with the descendant of another Dennis Tully meant his Dennis was son of my husband's second great-grandfather. So I constructed a "What if Not" scenario: what if Dennis Tully, husband of Margaret Hurley was not son of Denis Tully, husband of Margaret Flannery?

In fact, I drew up two proposals. For one scenario, I presumed that it was our Denis Tully's brother who was father of the other Dennis Tully, producing a proposed relationship between the two matches of fourth cousin. In the other proposal, I took a step backwards and played with the notion that our Denis' father had a brother who was father of the other Dennis, resulting in a relationship for the matches of fourth cousin once removed.

The results? Considering my husband shares forty five centiMorgans with this Tully DNA match, despite falling within the range given of shared cMs for each relationship possibility, there is a mere 11% chance that two people sharing 45 cMs would be fourth cousins. Worse, for the second scenario, the chances dropped to 4%.

And for the original proposal? The one in which the younger Dennis in Ontario, Canada, was son of our Denis from County Tipperary in Ireland, the percentage rose to 35%. Still not a robust percentage, but certainly better than the others.

With that, while I would have liked to see more solid documentation than that one report on Dennis' Canadian death certificate, I feel a bit more comfortable accepting that proposal. And just in time, too, for tomorrow we move on to another ancestor quest.


No comments:

Post a Comment

Related Posts Plugin for WordPress, Blogger...