Detecting and understanding the genetic basis of multivariate traits related to human health are critical to the future of personalized medicine. An important aspect is understanding pleiotropy (when a single gene influences more than one trait), which can improve the biological understanding of a gene in multiple ways, and ultimately advance prevention and treatment of complex diseases. However, the statistical methods to evaluate the simultaneous impact of a gene on multiple traits have mainly relied on standard multivariate analyses that do not directly address biological questions. We recently developed novel statistical methods to evaluate the association of a single-nucleotide polymorphism with multiple quantitative traits, overcoming the limitation of standard multivariate analysis in order to improve biological understanding of how a gene influences multiple correlated traits. We propose to build on our experience in order to develop new statistical methods that allow for different types of traits (e.g., binary, ordinal) in order to facilitate human genetic research, such as use of the rich medical diagnostic information from electronic medical records. We also plan to develop statistical methods that decipher the genetic basis of multivariate traits in the context of genetic pathway analyses. To enhance understanding of how genes influence multivariate traits, we plan to develop and evaluate causal mediation models in order to provide guidance on the most likely sets of models that ?explain? the association of genes with traits, thereby providing much needed guidance to epidemiologists and laboratory scientists on follow-up studies.
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.