Genetic ancestry and molecular epidemiology.

It is hard to believe that as recently as a decade ago, some of us were finding it virtually impossible to get an epidemiologic study funded from the National Cancer Institute if it required collection and analysis of DNA. Now, it is virtually impossible to get anything funded if you do not! The rapid transformation of epidemiology into the molecular world has been so swift and dramatic that by some metrics, we are still reeling from the impact and struggling to adjust. Having served as an Associate Editor for the journal for roughly 4 years and as Senior Editor for 1 year, I have seen this played out repeatedly in manuscripts assigned to me for review. Given this opportunity to write an editorial of my choosing, I decided to share some perspectives on where we often tend to fall short. Let me preface my comments with an acknowledgement of my biases. Although my doctorate is in epidemiology, I have post-doctoral training in statistical genetics under the tutelage of Robert C. Elston. I have also been involved in genetic epidemiology long before it became fashionable to do so—back before a single major cancer susceptibility gene had been cloned. In fact, I remember well the signal important discovery of inherited mutations in the retinoblastoma gene (1) and therefore the proof-of-principle of Knudson’s two-hit hypothesis (2). Thus, my bias is about the importance of an understanding of genetics and the fundamental laws of inheritance that is an essential prerequisite to molecular-genetic epidemiology. Consequently, I would like to share some simple suggestions designed to help improve the quality and relevance of our work.