B1944
Title: Inferring a directed acyclic graph of phenotypes from GWAS summary statistics
Authors: Tianzhong Yang - University of Minnesota (United States) [presenting]
Rachel Zilinskas - Statistics and Data Corporation (United States)
Wei Pan - University of Minnesota (United States)
Xiaotong Shen - University of Minnesota (United States)
Chunlin Li - Iowa State University (United States)
Abstract: Estimating phenotype networks is a growing field in computational biology. It deepens the understanding of disease etymology and is useful in many applications. A method is presented that constructs a phenotype network by assuming a Gaussian linear structure model embedding a directed acyclic graph (DAG). Genetic variants are utilized as instrumental variables and show how the method only requires access to summary statistics from a genome-wide association study (GWAS) and a reference panel of genotype data. Besides estimation, a distinct feature of the method is its summary statistics-based likelihood ratio test on directed edges. The method is applied to estimate a causal network of 29 cardiovascular-related proteins and is linked to the estimated network of Alzheimer's disease (AD). A simulation study was conducted to demonstrate the effectiveness of the method.