Using the Parental Phenotypes in Case-Parent Studies

• Activity Number: 554
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): David M. Umbach* and Min Shi and Clarice R. Weinberg
• Companies: NIEHS/NIH and National Institute of Environmental Health Sciences and NIH/NIEHS
• Keywords: case-parent triad ; parental phenotype ; association study ; SNP ; log-linear model ; log-binomial model

Abstract

Studies of case-parent triads often collect, but typically ignore, information about parental phenotype. Our approach uses the parental phenotypes to assess association independently of the usual case-parent-based association test, thereby enabling cross-generational internal replication of findings. If the disease under study does not interfere with reproduction, the method resists bias due to population stratification. We also combine information from the two generations into a single model that can exploit approximate equality of parental and offspring relative risks to improve power. When some parental genotypes are missing, one can fit the combined model via the expectation-maximization algorithm. We develop another composite test as a linear combination of test statistics from the parent-phenotype-based test and the traditional log-linear, transmission-based test. We evaluate the performance of our methods and compare them to previous proposals through non-centrality parameter calculations and simulations. We show that incorporation of parental phenotype often improves power to detect a genetic association with disease.