Functional and population genetic architectures of complex disease

Project Summary Recent progress in the genetics of complex diseases, including neuropsychiatric diseases, has revealed that the bulk of disease heritability is explained by the action of many common non-coding variants. Most of these variants exert individually tiny effects on disease risk, but collectively they contribute most of the measurable genetic risk for most diseases and typically account for between a third and half of twin estimates of heritability. These observations motivate three important questions concerning the biology underlying phenotypic variation: First, how do individual non-coding variants associated to disease impact biological function? A growing number of studies, including those published by the Key Personnel of this renewal application, provide statistical analyses that link genetic and epigenetic data to identify regulatory elements active in cell types that are key to disease biology. However, the mechanism of regulatory action is understood just for a few examples of genetic associations to disease. Second, how do the causative variants conspire together to perturb genes, biological pathways and networks and induce a disease phenotype? Despite the body of knowledge on genes and functional modules accumulated by decades of experimental research, we lack understanding of specific gene programs and networks through which the thousands of causative variants act to impact disease phenotypes. Third, how is genetic variation associated to common neuropsychiatric diseases stably maintained in the population given the loss of fitness associated to these disorders? Established population genetics models applicable to rare diseases are inconsistent with recent data on common disease genetics. We propose to develop new statistical and computational methods to generate biological insights from genomic data and to apply these methods to genome-scale genotype-phenotype datasets. We will design new strategies to combine functional genomic and molecular phenotype data with disease association results to shed light on the proximal regulatory function of non-coding variants. We propose new statistical methods to interpret genetic association data at different levels of biological organization ranging from individual regulatory interactions to genes, pathways and networks. We will use population genetics models to address conceptual issues of the origin of allelic architecture of common disease and, in particular, neuropsychiatric diseases. Our collaboration has an extensive publication record and a record of producing widely-used open-source software in the previous funding cycle. We have extensive computational and statistical expertise, but our approach is always rooted in data. All proposed method development will be guided by available large-scale genetics datasets and functional genomics datasets spanning over 2 million samples.