It isn't surprising that millions of people are interested in learning about their ancestry. I certainly am, so I assume most people are. For those who are new to this idea: numerous companies now offer (online, mostly, or sometimes in junk mail) to give you an analysis of your ethnicity; that is, where your ancestors probably came from, based on the genes in your DNA.
DNA is a protein-like molecule that lives in the nucleus of every cell in your body. All the DNA molecules in any one person are identical, and essentially unique to that person, unless he or she has an identical twin, or something like that. (Triplets, Quadruplets, and so on. If you're one of Octuplets . . . just kidding.)
It is a long strand, organized into 23 chromosomes, which we can think of as a list of 23 smaller lists. As far as I know--I'm not an expert--these lists, which are lists of genes, actually, together define how your body is built, and how it functions. For instance, in some one of these 23 chromosomes, in some specific place in its list (remember, each chromosome is itself a list!) there is a gene that specifies what shape your right ear is, for instance. Unless I'm mistaken, that gene, in that position, in anybody's DNA, specifies the shape of their right ear. Another gene specifies sensitivity to the sun, let's say.
Many genes have functions (they specify things) that are not known. Many characteristics (such as ear shape) might be specified by several genes. But the essential fact is that if two people have a segment of their DNA string identical, they will share some characteristic, or several characteristics, but of course, they might be very minor characteristics.
Now this whole string comparison thing falls plumb spang in the middle of the discipline of Mathematics, specifically in the area of analytical topology, so naturally I was curious about how they went about doing this ancestry business. And, to make it more interesting, they were reporting to their clients the places from where their ancestors probably came! How did they do that?
I went on the Internet, and without trying very hard, found this web page entitled
Pulling back the Curtain on DNA Ancestry Tests. Exactly what I wanted! Of course, I was not so naive as to expect a detailed description of the process, but it turned out to be a lot more useful than I had expected.
First, a warning. The article states that these companies get far more money by selling their information to commercial companies that can use the information (e.g. medical research outfits, or perhaps even companies that have less worthy objectives) than they get from you. So the $100 or so that you pay for the service hardly compares with the money they get in other ways, and they might as well give you the service for free, for supplying them with a data point. Also, the privacy agreement they make with you will not be valid if, for instance, they sell the company, or it is acquired by another company somehow. So you have to be prepared for your genetic information to be compromised almost certainly. For instance, if your DNA string information falls into the hands of a company that is investigating a certain variant of a gene to see whether it is a useful indicator of some horrible medical condition, and if they discover that you have it, well, they're sort of morally obliged to let you know that you're at risk for this condition, and depending on the laws of your state or the country, disclosing this information to, say, a medical insurance company, which is obliged to raise your insurance rates! None of this might happen, but then, it might, especially in the very business-friendly political climate in which we try to live. (And think: the company that buys the DNA information might be a subsidiary of an insurance company in the first place, which could have reciprocal agreements with who knows how many other insurance companies? But most people will have nothing to hide, and I do not want to chill your possible interest in discovering the various skeletons in your ancestral ethnicity closet.)
From what I understand, these companies mostly check genes, in the sample you provide, for genes that are common in certain regions, and which are uncommon in other places. It all depends on how they decide that this gene (gene variation, actually; everyone has to have all the same genes, or they would not be human. Everyone has to have a WX67 gene of some kind, but that gene could come in different varieties. I just made up the name of that gene, but you get the idea. Some people will have WX67, variant 1, and others will have WX67, variant 2, and so on, all the way to variant 4796. Yes, there are thousands of variants, called mutations. Viruses, it seems, have the most types of mutations). For example, a company will get hold of a number of samples of DNA from native Americans from some tribe, and study the sample for genes that are common to all the samples, and particularly uncommon in almost any other DNA. (This is a big undertaking, and if the company takes this job seriously, they have to be given credit. Still, the idea is simple.) Now, if you send in a sample of your DNA (basically a little saliva), and that gene is present in your DNA, they would conclude that you have some ancestry within that tribe of native Americans.
If they have identified, say, 100 different gene variants in that tribe of native Americans, and you have every single one of those, then the probability that you have that type of native American ancestry is very high indeed. Honestly, it won't be necessary for you to have all 100 gene variants for them to declare with 100 certainty that you have that ancestry.
In addition to telling their clients that they have some ancestry: say, Middle Eastern ancestry, these companies tell their clients what percentage of their ancestry comes from that region. Again, we can only guess how they arrive at those numbers. They look at large numbers of genes (remember there are a vast number of genes, so they're still looking at less than x% of your genes to make these conclusions. But that alone is no reason to doubt the validity of the conclusions; the tests used could be very subtle and delicate indeed; we just don't know how hard they have worked on these tests, and we have to trust them. So, we have to take the results of these tests with a dash of ketchup. Furthermore, there is the possibility for checking sets of genes, which will provide a more subtle analysis, a finer sieve of results. (Maybe that last observation was redundant; we were talking about the possibility of 100 different genes in a given tribe, after all.)
By this time, you have to have concluded that I'm trying to explain something that I'm not entirely qualified to explain, and you would be right. But you would certainly have some information on which to hang your reasoning now, and against which to assess the documentation they provide, if you choose to go through with the test. Remember, there are several companies you can get the service from.
No matter what you do, you should think clearly about your potential heredity. If you're black, you could still have Viking ancestry, because at least a few of the plantation owners could have been from the Scandinavian countries, and they were known to make free with the female slaves they owned. (However, it is also well known that not all plantation owners were cruel to their slaves, though the ones who were humane were probably largely in the minority.) You could watch YouTube videos of those who got the service, and see their reactions. The responses are initially surprise, but in hindsight, every instance that I saw was absolutely plausible.
Arch