Linking family trees and DNA

You might, if you were so inclined, call it the Holy Grail of genetic genealogy.

Putting your tree online with your genetic information, and having it link — automatically — to everyone else you’re related to.

Imagine, for a moment, being able to quickly identify the man who has to be your direct paternal ancestor, just by entering your DNA data with your personal data. Or being able to quickly identify the woman who is the furthest back in your maternal line.

It’s exciting stuff, this tie between family trees and genetics.

It’s what AncestryDNA is trying hard to do with its tree-matching function and its autosomal DNA tests. There, it’s exceedingly problematic, because (a) autosomal DNA changes with every generation so you can start seeing gaps in the genetic data by the third cousin level, with as many as 10% of third cousins not having enough autosomal DNA in common to show up as a match¹; and (b) autosomal results can be skewed by cousins marrying cousins as well as by the random nature of the way autosomal DNA is inherited.²

And it’s what WikiTree is trying now to do with its newly-announced DNA Ancestor Confirmation Aid, and in a much more useful way.

WikiTree, you may know, is a free crowdsourced effort “to grow a single worldwide family tree that will make genealogy free and accessible for everyone.” Members “privately collaborate with … close family members on profiles of modern people” and since “all the profiles are connected on the same system the process is creating a single family tree that will eventually connect us all and thereby make it free and easy for anyone to discover their roots.”³

This new DNA Ancestor Confirmation Aid lets every WikiTree contributor indicate whether he or she has taken a YDNA test (showing the direct paternal line and requiring a male descendant to test⁴) or a mitochondrial DNA (mtDNA) test (showing the direct maternal line of the test-taker, who can be male or female⁵).

It then provides a neat way to determine who else may be on WikiTree who might share your YDNA or mtDNA. It shows you five generations of your ancestors, back to your third great grandparents, and lets you quickly and easily see who else is on WikiTree who descends from your ancestors. If any one of those people has tested, there’s a tool that lets you compare the results and to indicate whether there’s a match as you expect.

And if nobody has tested, well, you now know who to go after for a sample, don’t you?

It’s a very nice addition to the genetic genealogy toolbox, especially — as noted by Roberta Estes of DNAeXplained, for folks who have mtDNA tested for whom there is no other real connection system.⁶ Melding and blending genealogical data and genetic data can be terrific.

But even with the WikiTree system, it’s necessary to add one cautionary note:

Yes, it’s where we all want to go.

And it is oh so very easy to be dead wrong.

Here’s the hitch. Let’s say John Doe, a WikiTree member, has YDNA tested and he enters his information into his profile. What WikiTree then does with that information is this: “If a male WikiTree member has taken a yDNA test we go up the male line, father-to-father, to find the member’s earliest-known paternal grandfather. Then, starting with this earliest-known paternal grandfather, we go down the various direct male lines. The system includes all sons and skips all daughters and profiles without a specified gender.”⁷

If John has mtDNA tested, it does the same thing for the direct female line: “If a WikiTree member (male or female) has taken a mitochondrial DNA test we will go up the direct maternal line and then back down all appropriate lines and attach the test information to their profiles.”⁸

Once John enters his information, this is the statement that appears on the page for every single person to whom his DNA result is attached by this automatic system: “It may be possible to confirm family relationships with Ancestor by comparing test results with other carriers of his Y-chromosome or his mother’s mitochondrial DNA.” That’s good. And then it says that there are either “Y-chromosome DNA test-takers in his direct paternal line on WikiTree” or “Mitochondrial DNA test-takers in the direct maternal line”.

And that’s not quite so good.

Because John Doe could be wrong.

If John’s tree is wrong, if there’s a gap in his documentation, the people he identifies as being in “his direct paternal line” or “direct maternal line” may not be his ancestors at all. And yet every single one of those people, starting with the very earliest ancestor John identifies as being his ancestor (Charlemagne, for example…), is going to have John’s DNA information attached to his or her profile.

Only when someone else who identifies himself or herself as a descendant of one of those people also DNA tests and takes the time to compare results and determines that there is a mismatch is there going to be so much as a hint that John’s DNA linkage information is wrong. And even then it requires member action to figure out where the error is and correct it, in the trees and not in the DNA linkage. WikiTree explains:

It’s important that non-matching tests don’t appear on an ACA and an ancestor’s profile indefinitely.

They remain there for as long as the incorrect relationship(s) that connected them remain on WikiTree. That is, we don’t edit DNA test connections directly. We edit WikiTree relationships. Test connections are automatically rebuilt at night based on those relationships.

It might seem easier just to remove the tests. However, it’s part of our community ethos that we are all striving for an accurate single family tree. Wiki Genealogists want to remove inaccurate relationships from WikiTree. DNA testing is a powerful tool for identifying these inaccuracies.⁹

WikiTree acknowledges that — even at its best — its system involves something less than proof: “Confirmation is not absolute proof. There are always caveats. For example, a man can be an exact yDNA match with the man who raised him, yet his real father could be the man’s uncle or the man’s grandfather, etc. Scientifically speaking, absolute proof of a theory is always impossible. A theory can be tested — a genealogical theory can be DNA tested — and the test can confirm or disprove the theory. A test can definitively disprove something. But a test cannot definitively prove anything, strictly speaking.”¹⁰

So while this is a good tool, and it’s a good step forward, we need to recognize its limits.

The fact is, on WikiTree, on AncestryDNA, on any system where we’re trying to blend genealogy and genetics, we need to understand the risks of what’s called GIGO: Garbage In, Garbage Out.

And, if we join in this crowdsourced work, we need to make sure our own work isn’t garbage.

SOURCES