A Mixed Linear Model Method for Association Testing on the X Chromosome in Samples with Unknown Structure

• Activity Number: 561
• Type: Contributed
• Date/Time: Wednesday, August 12, 2015 : 10:30 AM to 12:20 PM
• Sponsor: WNAR
• Author(s): Caitlin Mchugh* and Timothy Thornton
• Companies: University of Washington and University of Washington
• Keywords: statistical genetics ; population structure ; GWAS ; mixed linear models ; relatedness ; admixture

Abstract

Genome-wide association studies (GWAS) with millions of genetic markers aim to identify genetic variants involved with human traits. Statistical methods for GWAS have primarily been developed for markers on the autosomal chromosomes, with less attention given to the analysis of X chromosome markers. Mixed linear models (MLMs) are powerful and effective for analysis of GWAS with population and family structure. Existing MLM approaches are not directly applicable to analyzing X chromosome markers. We propose the MLM-X method for association mapping on the X chromosome in samples with population and/or family structure. MLM-X includes a random effect for polygenic effects on the X chromosome and random effects for the autosomes and environment. Variance components of the random effects in MLM-X are calculated via restricted maximum likelihood with empirical genetic relatedness matrices for the autosomes and X chromosome. We demonstrate that MLM-X improves over existing approaches, in terms of type I error and power, for trait mapping with X chromosome variants. We present MLM-X applied to a large study of Hispanics with genome-wide data and a variety of clinical outcomes and measures.