What DNA Can Tell You, and What It Can’t
The two most common types of genealogical DNA tests look at the DNA of the Y-chromosome (Y-DNA) and the DNA of the mitochondria (mtDNA). Y-DNA is patrilineal, passed down through father-line ancestry, while mtDNA is matrilineal, and is passed down through a person's mother-line ancestry. Men and women can undergo mtDNA testing, but Y-DNA testing is only an option for men because women do not have a Y-chromosome.
As it turns out, DNA testing can answer extremely specific questions and extremely general ones. Here are some of the scenarios in which it can be useful:
1. A specific question: "Are we cousins?"
Two people who suspect they have a common ancestor may be able to find out using a DNA test. Both individuals must be tested, and their results compared; if a significant number of identifying "markers" match up, those people are genetically related. Differences in the frequency of mutation of Y-DNA and mtDNA come into play. Since Y-DNA occasionally mutates as it is passed down from father to son, matching markers can provide only a probability of relation, although near-matches are considered useful; since mtDNA almost never mutates, perfectly-matching markers are much more conclusive, and near-matches aren't very useful. The number of matches reveals the probability that the two people share an ancestor within a certain number of generations. This ancestor is called the "Most Recent Common Ancestor," or MCRA. As explained at Wikipedia ("Genealogical DNA test"):
Y-DNA testing results are normally stated as probabilities: For example, a perfect 12/12 marker test match gives a 90% likelihood of the most recent common ancestor (MRCA) being within 23 generations, while a 67 of 67 marker match gives the same 90% likelihood of the MRCA being within 4 generations back.
2. Another specific question: "Are our families related?"
If your last name is Crow and you're from Gastonia, North Carolina, are you related to other Crows in the area? What about the Crowes? And if the Gastonia Crows are related to the Gastonia Wolfs, are they related to the Wolfes up the road in Charlotte? As with the cousin question, you'd need to get others to participate in such a surname comparison.
If you're using a DNA test to find specific answers—data that one could enter into a family tree—you're dependent on others' test results. Kimberly Powell, About.com's genealogy expert puts it:
An individual's DNA test provides little information on its own. It is not possible to take these numbers, plug them into a formula, and find out who your ancestors are. The marker numbers provided in your DNA test results only begin to take on genealogical significance when you compare your results with other people and population studies. If you don't have a group of potential relatives interested in pursuing DNA testing with you, your only real option is to input your DNA test results into the many DNA databases starting to spring up on the Net.
Results of a Y-DNA and mtDNA tests will identify an individual's haplotypes, which correspond to his or her haplogroups. A haplogroup is a group descended from a person who first exhibited a unique mutation in DNA. Haplogroups bind populations genetically but also by geography. American Indians commonly belong to the haplogroups A, B, C, D and X, which may have migrated from Africa to North America along the routes shown in the map above.
3. A general question: "What is my ethnic makeup?"
Some testing services offer something called autosomal DNA testing, which they claim provides geographic origins or biogeographic ancestry. This is the ethnic or geographic breakdown you may have seen on genealogy TV shows and, somewhat infamously, on George Lopez Tonight:
The test Lopez uses divides all of humanity into just four groups, and has its flaws (for instance, it found archetypal New York Jew Larry David to be 37% American Indian—but that's a story for another day). Other autosomal tests provide more granular breakdowns. But not all geneticists are sold on the accuracy of the autosomal testing options currently available. First, the tests are offered by private, for-profit companies that have developed their own databases and categorizations. Secondly, autosomal testing can blur the line between proportion ("you are 30% European") and probability ("people of European descent are likely to share 30% of your autosomal data")—anyone interested in taking an autosomal test ought first to brush up on some basic principles of statistics. As Roberta Estes put it in her analysis of DNA testing methods (PDF: "Successfully Using Autosomal Testing in Conjunction with Mitochondrial and Y-Line Testing to Address Genealogical Questions"):
Ironically, the results may vary significantly between [different autosomal tests]. There is no “right” or “wrong” answer at this point. I encourage everyone to simply view these results as “data”, hints to puzzle pieces. As the data bases improve and we better understand population migration and movement, the clarity of the results will improve too.
For genealogical purposes, even the most accurate autosomal data may be of limited use. Your ethnic makeup doesn't necessarily tell you anything about the identities or lives of specific relatives. If an autosomal test identifies a person as 12.5% African, it could be that the person has a full-African great-grandparent. But it could also be that there are any number of partial-Africans on both sides of the family tree, and there's no telling how far back. And in fact, there's no telling how much those partial-African ancestors knew (or wanted known) about their own roots. It's interesting to know that one is 12.5% African; it may not be possible to determine, based on historical documents, why.
In a recent article posted to BBC.co.uk, renowned genealogist Megan Smolenyak described autosomal testing as "an amusing test, but ... one should take the results as more of an indication than an absolute."
There is a key difference between genealogical DNA testing and traditional genealogical research. In traditional research, you can examine documents (letters your grandmother saved) and discover facts you were previously unaware of (I had a great-uncle Ray?).
But in order for DNA test results to be useful, you need to know what you're looking for. There are answers to be found—but it's up to you to ask the right questions.