Genome-Wide Examination of Common and Rare Variant Associations with Alcohol Response in the UCSF Family Study

Heritability estimates suggest that approximately 50% of the variance related to alcohol use disorder (AUD) can be attributed to genetic influences. Nonetheless, molecular genetic efforts to identify individual variants that... [ view full abstract ]

Heritability estimates suggest that approximately 50% of the variance related to alcohol use disorder (AUD) can be attributed to genetic influences. Nonetheless, molecular genetic efforts to identify individual variants that contribute to AUD have generated mixed findings, partially attributable to AUD phenotypic heterogeneity. Specifically, if differences in symptom presentation or comorbid conditions among individuals result from separable risk mechanisms, including such phenotypic heterogeneity will reduce power to detect genetic influences underlying these distinct mechanisms. One approach for addressing this heterogeneity is to study an endophenotype focusing on a specific facet of AUD (e.g., level of response [LR] to alcohol). LR is the extent to which a specific dosage of alcohol produces the responses associated with alcohol consumption. Notably, previous molecular genetic studies of LR using the Self-Rating of the Effects of Alcohol (SRE) questionnaire have reported associations with multiple variants and genes of interest, but due to small sample sizes have yet to be replicated. The current study is an initial effort to examine the effects of genome-wide common and rare variation in relation to LR to alcohol by utilizing low-coverage whole genome sequence data in a subset of the UCSF Family Study (N=1,188) with both genotype and SRE data for their first five drinking episodes. A GWAS of common variation (MAF>0.05) was carried out using the software package EMMAX. Gene-based tests of rare variation (MAF<0.02) restricted to coding regions were conducted utilizing SKAT-O. The GWAS resulted in no variants reaching genome-wide significance (p=5.0x10^-8). The top association was in an intergenic region of chromosome 3p26.1 (rs112429089, p=5.69x10^-8). Individual variants in multiple genes previously implicated with alcohol outcomes were nominally associated with SRE scores (e.g., TMEM132D, p=5.00x10^-6; CHRM2, p= 2.08x10^-5). For the gene-based analyses of rare variation, the top signal was an association with the SETX gene (p= 2.08x10^-7), with additional genes of interest being suggestively associated (e.g., ADCY4, p= 1.42x10^-4; BDNF, p= 2.26x10^-4). These findings suggest that similar to AUD, molecular genetic studies of alcohol response will require larger samples to identify specific risk-conferring variants, as well as alternative modeling techniques for investigating rare variation occurring at lower population frequencies.