Michael Cooley's Genetic Genealogy Blog GEN • GEN
8 May 2016

Evaluating SNPs and STRs Together

In the first of my present round of posts, I described the difference between Y-STRs and Y-SNPs. I'll start with a brief review. Some definitions:

  1. Nucleotides are simple sugar-based molecules. Genetic data is arranged in strings of four nucleotides, their names abbreviated to A, C, T, and G.
  2. Single Nucleotide Polymorphisms (SNPs) are just that, each being one nucleotide that has morphed, or mutated, to one of the other values.
  3. Short Tandem Repeats (STRs) are short strings of nucleotides that repeat in tandem to one another. For example, at DYS449 I have 34 repeats of TTTC.

SNPs and STRs exist throughout the genome. The mutations we find in mitochondrial DNA are SNPs. We see the terms often preceded by "Y-" simply to denote that we're talking about mutations on the Y chromosome, which is about all I've talked about in the last several posts.

STRs tend to be fickle. The count can fluctuate—up or down—from one generation to the next. Yet they are considered to be stable enough that general predictions can be made about relationships between two individuals going back approximately fifteen generations. That's not to say that precise predictions can be made, just enough that test results can be grouped together, especially when the surnames are the same. The comparisons become more significant when geographic regions and eras are also equivalent. Given that, I was surprised that the Hackett Y-STR match to the CF01 Cooleys is as close as it is considering that SNP testing has revealed that the common ancestor lived about eight hundred years ago—perhaps even further back according to some estimates. (I wouldn't be surprised if it turned out to be far more recent.) Still, no matter the number of hundreds of years involved, STRs can be remarkably stable—but that's not to be counted on.

SNPs are not so fickle. Once a mutation appears (a T flipped to a G, for example), it's passed down from one generation to the next unchanged. Stable SNPs have been found that are over 300,000 years old. With these, we can make much better predictions, such as the case with the Hacketts and Cooleys—from a highly nebulous relationship to one that some kind of year can be attached, even if approximate. But it would be a mistake to throw out STRs once we have advanced SNP testing done. Both are important pieces of data and can be used in tandem to gain a fuller picture.

There are three CF01 Cooley testers who are undoubtedly related but don't fit in the tree of John Cooley (c1738-1811) of Stokes County, NC, even though one bio says the family had origins in North Carolina. So far, the line doesn't appear on the radar until the birth of William Henry Cooley in 1797 and of James Cooley in 1808 or '09, both born somewhere in Pennsylvania. They were of the right age to have been grandchildren of John's but we can place none of his sons in Pennsylvania. Besides, each of the three descendants have two Y-STR markers that are not found among the Stokes County Cooleys: DYS464b has 15 repeats (as does the Hackett tester, which means that it is relatively old) and DYS576 has 18 repeats as opposed to the 17 repeats the rest of us have, including the Hackett tester, perhaps indicating that the extra repeat occurred more recently.

I've mapped both SNPs and STRs in this graphic, arranging them according to their relative ages. The "other" Cooleys appear down the far left arm. Not enough data is known to determine the order in which the DYS390 (STR), YP4491 (SNP), and CDY (STR) mutations occurred but the presence of the two latter STRs make it likely that the two lines split somewhere after the appearance of YP4491. Therefore, the two men born in Pennsylvania were probably of a collateral line rather than having descended from John—perhaps from a brother of his—but time will tell!

We can also throw Family Finder (autosomes) into the mix. The STRs for the Agerton tester (bottom) match the CF01 Cooleys—more so to the North Carolina group—and it's known that Agerton was an adopted name. He hasn't had advanced SNP testing but he does have an interesting Family Finder match to a descendant of Perrin Cooley, Sr. born in Stokes County, NC to John Cooley. It's not proof positive but it makes sense when taking into consideration that he has matches via two testing methods.