Statistical Methods for Association Studies of Complex Genetic Disorders

Jinko Graham and Brad McNeney


Research in the group is at the interface of statistics and genomics. At this interface, many analytic challenges arise from issues connected to populations or to study design. In particular, data with high dimension and complex structure present opportunities for statistical innovation in inference and computation. These opportunities include, for example, inference for data with complex dependencies at both the level of subjects and the level of the genome in the population, missing or partially observed data, and high-volume testing of statistical hypotheses. Despite the incomplete understanding of these issues, human genetics continues to advance quickly, propelled by the technology into uncharted areas such as whole genome sequencing studies and the study of gene-environment interactions using population cohorts. Our focus is on developing analytic tools to uncover patterns in such data, while accounting for random variation. Current projects include: (i) Bayesian inference of gene genealogies from genomic data, with application to the study of complex disease traits, and (ii) improved inference of gene-by-environment interactions, to understand how the health effect of an environmental factor is modified by genetic background in the population. In our collaborative research efforts we are working with experts in genetics and oncology, pediatric rheumatology and neuroimaging to understand the genetic architure of associated traits.