DNA and Double Cousins

 

Thanks to the early 1800s marriage of two Ambrose sisters and two Flowers brothers, several of their descendants among my husband's distant cousin DNA matches show up as far more closely related genetically than genealogically. In other words, according to DNA test results, some distant cousins share more centiMorgans than would be expected for that level of relationship. In one case in his family, the numbers are off the charts.

That is not an uncommon experience, as I've learned. Even here at A Family Tapestry, reader (and blogger) Lisa mentioned that double cousins in the family tree gave her a "boosted DNA relationship." In her case, she was referring to actual double first cousins, descendants of her grandparents and their siblings. In my husband's case, though, Elizabeth Ambrose's marriage to Joseph Flowers produced children of his third great-grandparents. With their double first cousins—children of Elizabeth's sister and Joseph's brother—we need to look to the most common recent ancestors among my husband's fourth great-grandparents.

Last week, I worked on the ThruLines results for descendants of Elizabeth and Joseph, our direct line. This week, I'm starting the same process for the descendants of Elizabeth's sister Susannah Ambrose and her husband, Joseph Flowers' brother. While of Mathias Ambrose's ThruLines DNA matches at Ancestry.com Elizabeth yielded us thirty five matches, of Susannah's descendants, we have a smaller number: seventeen matches.

That, of course, represents what Ancestry could find for me through the ThruLines process. Keep in mind, while ThruLines is drawn from DNA test results, it also utilizes information from other subscribers' family trees. If the trees are incorrect, the ThruLines readouts will obviously be less helpful. As Nicole Dyer of Family Locket puts it, "ThruLines is a computer algorithm, not a genealogist." She sees ThruLines readouts as hypotheses, not documented fact.

I like how Roberta Estes puts it: "ThruLines is a tool, not an answer." Thus, I tiptoe very carefully through the handy diagrams which seem to assert exactly how my subject (in this case, my husband) is related to any given DNA match.

In the case of Susannah Ambrose's descendants, that was a necessary precaution. For one DNA cousin, the ThruLines diagram leading through the six generations subsequent to Susannah included boxes which all had dotted lines around each name. That signified that none of those generations were included in my own family tree. Odd, considering that I already have most of Susannah's descendants documented in this family line—until I realized that the entire sweep of that family line belonged not to Susannah's son, but to a son of Elizabeth. Apparently, enough trees on Ancestry disagree with my assessment to convince ThruLines that Susannah was the correct ancestral mother, not Elizabeth.

More puzzling than that was the ThruLines readout for another descendant of Susannah. The first glitch which caught my eye involved the son of Susannah's great-grandson, Robert Henry Dittoe. Robert was born in 1866—at least, according to the date inscribed on his headstone. I'm not sure why, but the information on ThruLines for Robert's son indicated that his date of birth occurred in 1797. Obviously, something is seriously wrong here.

Despite that glitch, the descendant who called that man her grandfather turned out to be labeled as my husband's fifth cousin, once removed. That may make sense, considering the ancestor in common for them was Mathias, father to both Elizabeth and Susannah. But here's the problem: these two DNA cousins share over one hundred centiMorgans.

If you look at any chart on the odds of that happening—for instance, the Shared centiMorgan Project update at DNA Painter—you'll see that by the level of fifth cousin, that chance approaches zero. But this relationship is fifth cousin once removed. Hmmm. Unexpected parental event? I don't think so. I say the finger points more accurately at those double cousins lurking in the background.

Fortunately for those of us researching such DNA scenarios, statistical help is on the way. Over two years ago, The DNA Geek, (aka Leah Larkin), posted a plea for volunteers to upload their data for a project to be done in partnership with the DNA Doe Project to examine real results for descendants of double first cousins. The goal was to produce something helpful for those dealing with DNA test results from either endogamy or pedigree collapse. The response to that plea yielded enough volunteer participation to do a test run and develop a tool, the user interface for which is just now about to be tested in a beta program this month, according to a comment following the original blog post.

In my husband's case, this is one of the more extreme examples of unexpected numbers regarding DNA matches, but there are signs of pedigree collapse elsewhere in that family tree. All the more reason to anticipate the results of that beta test and learn how to put that tool to good use. Whether the results can tease out a pattern in how double cousins and pedigree collapse impact DNA test results remains to be seen. But from what I see of our own family's DNA matches, those are the type of results which do not fit any other mold.