ven a brief perusal of current journals and meeting programs shows the vibrant life and potential of in the 21st century. Even if no new epidemiologic methods or technologies were developed, we could still make major contributions to understanding the etiology of the many diseases that are underresearched, or are still poorly understood despite substantial research, or are newly emerging. There is still plenty to be achieved with the well-designed questionnaire, the optimization of response and follow-up rates, and the 2 X 2 table. However, this is also a time of rapid development of technologies that will facilitate epidemiologic research. Epidemiologists have always been opportunistic in the adoption of new technologies for exposure assessment or outcome definition, and many new tools have already become available in the last 25 years. In the early 1980s, Perera and Weinstein' extended the term molecular epidemiology from the study of characterization of infectious organisms to the study of exposure and suscep tibility relevant to cancer and other chronic diseases. For some exposure-disease rela tions, methods of exposure assessment have identified the etiologic culprits in a rogues' gallery of suspects-such as the association of human papillomaviruses with cervical cancer, after observational studies had strongly implicated aspects of sexual behavior as increasing risk, and seroepidemiologic studies had suggested associations with a number of other sexually transmitted infections. Similarly, the identification of somatic mutations characteristic of exposure to aflatoxin in the TP53 gene in liver tumors provided a crucial piece of evidence to bolster a previously disputed set of ecologic and dietary studies associating aflatoxin exposure with liver cancer risk.2 Application of novel biomarkers as measures of exposure, or to assess the misclassification associated with other methods, will offer new opportunities to define exposure-disease relations. Much of the rest of this article discusses the scale and accuracy of new methods to assess between-person inherited genetic variation. However, we (and the people who fund our work) should not forget the need to better assess environmental exposures, because these are often modifiable and remediable, unlike our inherited genotypes. In the last decade, the development of individual biomarkers has been revolution ized by -omics approaches. The suffix -omic is from the Greek meaning or every; thus, genomics is the study of all of the genes in an organism. The first genomic technology in widespread use was the expression array, which enabled the study of the degree of expression of all known genes in an organism (sometimes referred to as the transcriptome). The major application in large-scale human studies has been the analysis of tumors. The ability to measure the transcriptome of individual tumors permitted the description of different patterns of variation in gene expression within tumor types previously thought to be a single type based on conventional histology. These patterns have been shown to be associated with prognosis3 and treatment response.4 For some tumors, the different patterns may derive from different cells of origin.5
Read full abstract