The Annoyed Man wrote:Anyway, the DNA test worked as advertised, but it didn't really tell us anything you can't tell just by looking at her fair skin, freckles, blue eyes, and her brownish auburn hair color.
Yeppers. I haven't been active in genealogy for a couple of years, but the subject of DNA testing for that purpose gets complex pretty quick, and that's why I shot WildBill a PM offering a telephone call for a quick overview and brief Q&A. I became involved in it in 2004 and dabbled quite a bit over the span of a decade.
There are three types of DNA testing done for genealogical purposes: yDNA (the male Y chromosome, passed down only in the patrilineal line), mtDNA (mitochondrial DNA: which isn't human DNA at all but that of the mitochondria, those little cells none of us can live without, that are passed down through the matrilineal line), and autosomal DNA: the 22 pairs of chromosomes--minus the X and Y sex chromosomes--that make up the human genome. This is sometimes abbreviated as atDNA, sometimes as auDNA.
Choosing what and who to test starts with defining "why." The autosomal testing has gained a lot of popularity over the past three or four years, and is making a nice revenue stream for companies like Ancestry.com (to my knowledge, there are five companies offering autosomal testing in any volume, and one of those is the Genographic Project, a genetic anthropology study funded mostly by National Geographic and IBM).
One of the reasons for the popularity is, I believe, that while autosomal results are extremely complicated to evaluate, the most (monetarily) successful companies doing it do
not provide you with marker results to the gene level, but merely an attractive pie chart representing how those genes have combined over the generations to give you a fairly accurate breakdown of ethnicity--slash--anthropological diasporic regions of origin. They do this by using massive databases comparing known-genomic values (many of which stemmed from tests of endogamous populations by Dr. Spencer Wells that helped lead to the founding of the Genographic Project in 2005) to your values and arriving at mathematical probabilities about what matches and what doesn't.
The Ancestry.com test, for example (which I've taken), provides no access at all to the raw data resulting from your test, only the estimated ethnicity/anthropological breakdown, plus any presumed matches from others who have also taken their particular test and whose information is in their database. And that's one place it breaks down for people like The Annoyed Woman: recent relative matches are unique to each testing company (for the most part). If I take Family Tree DNA's autosomal test and TAM takes the one from Ancestry.com, we could be first cousins and never know because the data isn't translated from one company to another.
And because of the complexity of autosomal DNA (there are 22 of those chromosomes, remember, and Chromosome 1 contains some 2,800 genes, Chromosome 2 about 750, and so on; add to that the fact that genetic recombination when insemination occurs is not a clean 50/50 thing: it's more of a "one from the mother, two from the father, three from the mother" sort of thing and with certain dominant genes winning out over recessive ones), the matching of near-relationship accuracy diminishes very quickly. As you'd expect, confirmation of parent/child/sibling relationships is extremely accurate; first cousins is pretty good; when you get even to second cousins, additional genealogical paper trail information will be needed to corroborate the match, and you may need to identify additional individuals for DNA testing to help seal the deal. You get beyond a second-cousin relationship, testing of multiple family members will almost certainly be required.
And I say the latter because, as wonderful as the age of the internet has been to genealogy, it's also introduced huge amounts of erroneous information into the wild that amateur genealogists will find online, decide it's a match with their family tree (without rigorous evaluation of source material and application of the "genealogical proof standard"), copy it into their own GEDCOM as fact, only for their information to be subsequently discovered by another researcher...and then it becomes like the old pass-it-around-the-campfire game where a simple sentence becomes entirely something else once it's whispered from person to person a dozen times.
DNA is a genealogy fact-checker no one had available at the turn of the century, and for that use alone it has tremendous value to family researchers. That said, however, if the goal is to verify existing genealogical information, I'd recommend the first item on the shelf be yDNA testing.
I've already lost all readers and am typing like a maniac which I told myself I wouldn't do, so I'll try to offer a summary and then shut up.
The male Y chromosome passes down only from father to son, and has only one actual function that we know of, a switch of sorts: if present, the baby will a boy, and if not, a girl. Women will be able to intuitively verify this statement: most of the Y-DNA chromosome, as far as we know, is junk. In fact, that's how it's referred to, as "junk DNA." But it's highly valuable to genealogists because it's a strict patrilineal chromosome transfer--there's no recombination--and there are well over 100 sets of markers, or STRs, that have been defined and for which generalized mutation rates have been identified.
If you and a suspected second cousin have interwoven family trees on the patrilineal side and you want to confirm if you share a common male ancestor, yDNA can do that for you with a high degree of accuracy in genealogical timeframes, and offer an estimate of the probability as to the number of generations to that common ancestor. In my own case, I'm descended from Welsh Quakers who moved into North Carolina circa 1770. The family included four boys and, through paper-trail research, I found three likely male relatives descended from three of the brothers. We each had yDNA testing done with the result of incontrovertible proof that that those three brothers were siblings, that we had almost certainly verified that their father was our oldest known male ancestor, and that the fourth brother's descendants were also almost certainly interwoven with our own family trees. One of those research Eureka! moments that can have you shout out loud.
In its wisdom, nature allows mitochondrial cells to pass to an embryo only from the mother. The father's got no say in this matter. Each of us, whether male or female, inherited our mother's mitochondria. So now we're looking at the matrilineal line.
Again, good news and bad news. The good news is that we have genetic markers to evaluate for the strict matrilineal line, and the expansion of the number of identified markers keep refining the accuracy of the results. The bad news--for genealogists researching family-tree relevant timeframes--is that mtDNA mutates significantly more slowly than yDNA. Current state-of-the art full mtDNA testing panels can provide roughly a 50/50 probability of a match at five generations, or about 125 years. And 50/50 can be an indicator, but not proof. A complete match on the more common HVR1 and HVR2 panels, which I've taken, put the 50/50 chance of sharing a common maternal ancestor out to 28 generations. Not terribly helpful for genealogy--though it can disprove a suspected relationship if the marker values are off--but very valuable for anthropological timeframes.
In somewhat the same way as databases have developed in autosomal testing that allows Ancestry.com to give a predictive estimate about your ethnic/ancestral origins, both yDNA and mtDNA results have been categorized into what are called haplogroups (not to be confused with hoplophiles or hoplophobes, which is a whole nuther non-genetic 2A matter). Haplogroups share certain identifiable genetic markers and can be broadly used to estimate anthropological history. Further refining haplogroups are "clades" and "subclades."
Haplogroup nomenclature has changed multiple times over the past decade, and the level of detail has become such that, on the yDNA side, there are no longer subclade or sub-subclade naming, but only identification by the most detailed marker for which you match. For example, on my patrilineal side, I'm R-L151, which aligns with family oral tradition that we migrated to America circa 1730 from Wales.
Remains fascinating stuff to me, even though I'm no longer what I'd consider an "active" researcher.
