Problem of "false positive" conclusions in genetic epidemiology: lessons from the leukemia cluster near the Sellafield nuclear installation.

As you may imagine, I feel quite privileged to deliver the first presidential address to this Society. Although somewhat overlooked, in recent years the evolution of the field of genetic epidemiology has been just as dramatic as the evolution of other fields of genetic research, and the formation of this Society is a timely development. In 1954, W.J. Schull and I published a little textbook entitled Human Heredity. We took turns writing chapters for it. He, for instance, wrote those very nice chapters on statistical genetics. We felt there should be in the book a chapter on “Genetics and Epidemiology”-the first such chapter that I am aware of-and to me fell the challenge of writing that chapter. Anyone with any doubts as to how greatly the subject of genetic epidemiology has advanced in recent decades needs only go back and consider the comparative simplicity of what I wrote some 40 years ago. Currently, genetic epidemiology presents two rather different faces. Face one is concerned with teasing the genetic component out of complex sets of family data. Here the objectives may vary from an attempt to link a marker gene to a phenotype, to an attempt to identify the existence of genetically caused variation in a phenotype by means of complex segregation analysis. Face two, on the other hand, is Concerned