Motivation
Much of the field of genetic epidemiology has focused nearly exclusively on the relationship between genetic variation and disease outcome. However, identifying important environmental factors, and their interaction with genetic variation, is key for explaining complex traits and predicting the risk of developing disease. Further, much of the exploration of the relationship between genetic variation and outcome has ignored genetic interactions and the possibility for epistasis. The lack of heritability explained, and lack of models predictive of disease outcome, have led to more discussion of the importance of identifying contributors to disease beyond pursuit of single genetic variant – single outcome analyses.
There are multiple research opportunities for identifying important environmental contributions to disease, from individual environmental exposures to the “exposome”: the totality of exposures of each individual over the life course. For some diseases there are known environmental exposures that contribute risk, such as smoking or UV exposure and cataracts. There are other disorders where identification of environmental contribution to risk and disease severity are very active areas of exploration and discovery, such as understanding the etiology of autism. Identifying important exposures contributing to outcome from large numbers of possible exposures and toxins is a challenge, and high throughput screens are an important approach as well as methods such as environment-wide association studies (EWAS).
It is also important to understand the relationship between genetic architecture and outcome by exploring interactions between genetic loci. Epistasis has been detected in model organisms such as drosophila, and there is evidence of epistasis contributing to diseases like Alzheimer’s, diabetes, and cancer . There is also evidence of epistasis between rare variants affecting complex diseases. Detecting epistasis in humans has been challenging due to issues including computational limitations, especially for interactions higher than pairwise; decreased power to detect interactions when compared to main effects; and difficulties with replicating significant epistatic models across heterogeneous samples. There are many opportunities for advancing methods for detecting gene-gene interactions.