Follow-up on DNA testing

In reading my most recent book, I came across a statistic related to the "milkman" problem I referenced in my last post. In this post, I noted that DNA testing for genealogists could potentially raise problems, especially in the future, with the increased use of sperm banks.

The book I am reading now is Dr. Tatiana's Sex Advice to All Creation by Olivia Judson (Metropolitan Books: New York, 2002). Before you jump to conclusions, let me explain that it is a natural history book that discusses genetics and the many ways that plants and animals have evolved to reproduce. Here is a quotation (p. 163-4) that directly relates to the milkmen:

"For example, in England, children usually get their last name from their father. Boys get something else from their father: their Y chromosome. Thus, if all living males bearing a particular last name are the direct descendants of one man, they should all have the same genetic markers on their Y chromosomes. In the absence of infidelity (or adoption), last names and Y chromosomes should match up. One study analyzed the Y chromosomes of men called Sykes, a name that first appears in written records about seven hundred years ago. It turns out that almost all the Sykeses investigated did indeed have the same markers on their Y chromosome, suggesting that most living Sykeses have the same distant ancestor. The rate at which females married to Sykeses were unfaithful (or adopted sons) over the period of seven hundred years is estimated to be 1.3 percent per generation."

Judson references: Sykes C. and B. Irven, 2000. Surnames and the Y chromosome. American Journal of Human Genetics 66: 1417-19.

That's pretty incredible if you think about it. Of course, I can think of any number of possible flaws with the study, and since I haven't read it directly, have no idea how accurate it might be. Still, a 1.3% infidelity rate seems astronomically low. Added up, though, it can create a huge possibility for biological error (never mind all the possibilities for documentary or research error).

Here's the math based on the 1.3% above: a 98.7% chance you are related to your father, 97.4% chance you are related to your paternal grandfather (but 98.7% for the maternal grandfather), etc. There is a 96.1% chance that you are biologically related to your grandparents... yes, the ones listed on your parents' birth certificates.

By the time you hit your ggg-grandparents, there is only a 59.6% chance that you are biologically related to all of them. The only person you can be sure of in this model is your mother's mother's mother's mother.

This means that there is an exponential rate at which you are likely to not be related to your ancestors. Based on the math, in under 2000 years, you are guaranteed to not be biologically related to your father's father's father...

Ironically, of course, there is also evidence that with the world's population size about 2000 years ago, we are all likely to be related somehow. Just not the way we think.

In another similar study, Y-chromosomes and the extent of patrilineal ancestry in Irish surnames, the authors note that: "Notwithstanding differences in their early origins, all surnames have been extensively affected by later male introgession." In other words, although we were all once 'pure' lines descended from one or multiple males (depending on the surname), we're mostly all bastards now! There was about a 50% likelihood of relatedness in this study between two males with the same surname. Not bad, but less than you'd expect without the 'introgession'.