Questions come up occasionally about my use of genetic genealogy: using genetic testing for matching family. You may have seen TV ads for 23 & Me or AncestryDNA tests. It’s becoming more and more popular, and I think it’s worthwhile to post a quick overview of what it is, what it can do, and what it can’t. Don’t worry, you won’t need a biology degree for this.
This is just the nickel tour… there’s lots of detailed posts out there that get into the science behind it all, but I thought an overview might be interesting for some of you.
The very short answer is that genetic genealogy an exciting tool, but one that enhances rather than replaces traditional genealogy. People can be adopted and not be aware of it, they can lie on official documents, and genealogy is full of the dreaded “NPE” – the non-paternal event, wherein the listed father is not the biological parent. Knowingly or unknowingly, there are mistakes in the written record, and, eventually, there are not any remaining written records.
Genes, on the other hand, can’t lie. They don’t often point a finger, either; most of the time, they wave a hand in the general direction. Still, it’s a handy tool for genealogy, and a unique one: while it can only be used by the living, results (in the form of matches) can appear for years to come.
The Four Kinds of Genetic DNA
In the Star Trek universe, we’d have full genome coverage at our fingertips. We’re not there yet. The original Human Genome Project, which ended in 2003, sequenced the entire human genome at a cost of $2.7 billion (with a “b”) dollars. Twelve years later, we’re starting to have companies that will sequence your genome for $1000. Not cheap, but definitely within the reach of enthusiasts.
Likely the price will continue to drop, but even as it does, it won’t be helpful until many people have decided to test. A genome without matches is just an interesting scientific experiment. So, pragmatically, what are the DNA test available to us today for purposes of genealogy?
You probably recall from science class that humans have 23 pairs of chromosomes, including one set (XX or XY) which determines sex. Autosomal DNA are the other 22, segments of which are inherited more-or-less evenly from the mother and the father. The inherited segments continue to split in subsequent generations; the size of segments with respect to other people who have tested their autosomal DNA indicates how recently you shared a common ancestral couple. Multiple people sharing the same segment are descended from the same line.
That’s greatly simplified. The reality is that smaller segments can stick around for a very long time, moderately-sized segements for more generations than were originally thought, and that endogamous populations – like Colonial America, island populations, and Ashkenazi Jews – are related along multiple lines, and thus appear to be more closely related when comparing autosomal DNA than do other populations.
Still, it’s the most accessible and most helpful test at this time for confirming recent genealogy. My experience with Ashkenazi kits is that it’s extremely accurate up to third cousins, and then drops off precipitously in differentiating 4th or 5th cousins vs. (say) 8th or 9th cousins. They’re all related, but in many cases well beyond the available documentation.
The Y-chromosome, which determines the masculine sex, can also be tested. Without going into the details here, the more common testing methodology are Short Tandem Repeats (STRs), which slowly mutate from generation to generation. Usually, YDNA is passed from father to son unchanged, but in one of an estimated 10,000 births, there are one or more mutations. As it happens, I have at least one mutation from my father.
Since different STRs have known average mutation rates, they are, in the aggregate, a pretty good indicator of the timeframe for a most recent common male ancestor, or for proving that there is no such person in the last few thousand years. What it cannot do is nail down who that ancestor is. For example, this test cannot determine the difference between two brothers, or two male first cousins, or any two men, if they descend from the same male lineage in the recent past.
My YDNA test connects my paternal line to the medieval Rabbinic “dynasties”, who were one of the few groups of Jews in the Middle Ages to keep good records, and thus their descendants tend to have high-quality trees. Of course, there were always those who “embellished” their family history by claiming a relationship, and who could say otherwise? Turns out, DNA can.
How my family went from Rabbis to blacksmiths by 1800 must be an interesting story; perhaps there’s an NPE involved. But I did visit the resting place of Rabbi Zevuun Eliezer Heilprin when I was in Jerusalem last year. Don’t quite know how we got here, great^14 (give or take) grandfather, but it’s nice seeing you….
Mitochondrial DNA, is passed from the mother to all children. mtDNA mutates even slower than YDNA, and thus is less useful when trying to coordinate with most written records. A perfect match using mtDNA suggests a 50% chance of a common female line ancestor within the last 600 years… and 50% chance that it’s even further back. Not my personal favorite test for that reason – which hasn’t stopped me from investing in such things, for what the future might bring.
The OTHER sex chromosome, passed from mother to child and, for daughters, one copy unchanged from the father. This is usually not a separate test, but falls out of the test for autosomal DNA. It was believed that comparison of X-DNA would provide some interesting help when combined with other tests, since (eventually) it comes only from female relatives. In practice, the X-DNA inheritance rules have stubbornly refused to be clearly identified. They don’t follow the 50% rule we see from other DNA, and in some cases children seem to get one of the mother’s X chromosomes intact, as opposed to some combination of the mother’s two X’s. Thus, it’s possible that two full siblings could have NO XDNA match. No one really knows how to interpret X-DNA yet, although if it’s there, there’s a relationship.
There’s three major companies doing DNA testing and matching: Ancestry.com (AncestryDNA), 23&Me, and Family Tree DNA (FTDNA). There’s several other companies that will TEST your DNA, but if you’re looking for matches in genealogy, you need comparisons, not just results.
AncestryDNA (where I tested first) probably has the largest data base, but thus far has declined to give good tools to look at your results. Instead, they “suggest” relationship levels. They do allow you to link your Ancestry.com tree, which is nice, and if your match also has a tree with the same people, they’ll be flagged. Thus far in five years, I’ve received no flags.
23&Me has been around for a while, but in the past has largely focused on testing for possible medical conditions. As such, it appears that matching testers there are less inclined to communicate, as they are not looking for new cousins. The company has more recently attempted to pivot towards genetic genealogy; hence, those TV commercials. Still, this is a real shakeup in their business model, and it’s difficult to recommend them at this time.
FTDNA, where I’ve done the majority of my testing, is the oldest company in this business (from 2000), and his heavily involved in scientific research to improve tests and interpretation. They have by far the best tools, allowing you to view matching DNA segments, download matches, and even your entire raw results if you are so inclined.
Where I am with all this….
I have nine test kits at FTDNA, including myself and both of my parents. In retrospect, I could have skipped myself, but I started with just me. “Hidesight is…” and all that. A couple of people were excited enough with the prospect, and in a position to test on their dime; the rest were very generous and tested at my request. Two of the people who tested have already passed away, including my grandfather’s second cousin, and I’m thankful I started no later than I did!
Each kit has roughly 5500 (!) matches – we Jews like to test, apparently – so there’s a lot of cross-referencing I do myself with Excel. So far, there have been two partial successes:
One is a perfect Y67 (YDNA at the 67 STR tested level) match with a gentleman whose family came from the Ukraine. That is, his direct paternal line and mine meet up somewhere. His great-grandfather’s given name was Boris (Baruch), and he was a blacksmith. My thinking is that he was a descendant of my gggg-grandfather Berko Fingerbren (~1800-1855), or possibly one generation back from that. It’s unlikely (although nothing’s impossibel) to be earlier with a perfect match. Additionally testing on that side would help somewhat.
The other success is an autosomal match with a JACOBSON family originally from the area of Lithuania from whence came my gg-grandmother, Frieda JACOBSON SCHNITZ/BLOCK. This match also cross-referenced as a match with my father’s cousin, who is also descended from the same JACOBSON line. We don’t have the paperwork connecting us, and that region of Lithuania is particularly bereft of surviving documents, but there’s a very strong connection nonetheless because of the DNA tests. The genes don’t lie.