A post by Manjunath Vadiari at Theories on past events provides a warning about taking the migration histories that testing companies provide along with genetic markers too literally. It's easier to sell a heritage than a set of numbers (this is what most people are paying for after all), but the history can be inaccurate, especially at low resolutions and when few subclades have been defined (especially common outside Europe). Manjunath, a Malayali living in Hyderabad, tested as Haplogroup R1 (M173), one of Eurasia's most common Y chromosome haplogroups, and got a "personalized" message from Spencer Wells describing his ancestors' journey from Central Asia to Europe. R1a is a very common haplogroup in places like Kyrgyzstan and Poland and R1b in Spain, England, and Ireland; the last two locations provide the Ys for a significant proportion of Genographic Project's R customers. The video form letter usually goes over well, but it doesn't quite work when it strikes an informed customer as profoundly unlikely.*
Apparently he got word back from Dr. Wells that instead he is R* (M207), a group ancestral to both R1 and R2, which is more common in South Asia. Still, be careful. This is an idea that has implications for all sorts of testing bias while data are slim. A Siberian D could easily get a story of her ancestors' settling of the Americas (really her very distant cousins') or an Ethiopian L3 of his forebears' journey out of Africa. Actually, nearly all of the literature focuses on what happened on the way out of Africa, rather than the very interesting things that happened *in* Africa (fortunately something that Wells himself is concerned about and working on). The results of your DNA test should be an evolving story**; its cladistics should show you a more complex (and thereby more accurate) story as you learn more and find closer and closer matches.
To say nothing about the limits of using only markers expressed in your *own* DNA, or of ascribing too much to *one* genetic identity... but I'll be talking about that forever, so it's not for this post.
* This is not the same as the "unlikely" of the uninformed customer who wouldn't immediately think of West African migrants to Britain becoming Yorkshire patriarchs as plausible, or Danish or Nigerien patrilineal descendants of Genghis Khan. Manjunath well knows that 35 kya Central Asians became ancestral to many migrations west, east, and south, and for a deep clade like R*, a move from Central Asia to Kerala via the Khyber Pass and Deccan is more likely than one that went all the way to Europe, then back.
I didn't mean that, seriously. The title came after the unintentional pun.