GENOME-WIDE INVESTIGATION OF ALCOHOL RESPONSE: A META-ANALYTIC REVIEW

Heritability estimates from twin and family studies suggest that approximately 50% of the variation related to the development of alcohol use disorder (AUD) can be attributed to genetic influences (Verhulst et al., 2015).... [ view full abstract ]

Heritability estimates from twin and family studies suggest that approximately 50% of the variation related to the development of alcohol use disorder (AUD) can be attributed to genetic influences (Verhulst et al., 2015). Nonetheless, molecular genetic efforts attempting to identify individual variants that confer risk for the development of AUD have yielded few replicable results suggesting that large sample sizes achievable through meta-analysis will be required to identify robust associations. An additional challenge in the genetic study of complex traits such as AUD is the substantial phenotypic heterogeneity found within the AUD diagnosis. One strategy for addressing this heterogeneity is the examination of AUD endophenotypes that focus on specific subfacets of AUD (Gottesman & Gould, 2003). One well-established AUD endophenotype is level of response (LR) to alcohol. LR is the extent to which a specific dosage of alcohol produces the responses typically associated with alcohol consumption (Schuckit, 1980). One widely-used retrospective measure of LR to alcohol is the Self-Rating of the Effects of Alcohol questionnaire (SRE; Schuckit et al., 1997). Notably, previous molecular genetic studies of LR, including studies of the SRE, have reported associations with multiple variants and genes of interest; however, these studies have also resulted in inconsistent findings (e.g., Joslyn et al., 2010). The current study is an initial effort to extend results from previous molecular genetic studies of LR to alcohol by conducting a meta-analysis of two genome-wide association studies (GWAS) of SRE participant reports of their first five drinking episodes (SRE First 5). GWAS were carried out examining common variants (MAF>0.05) for each sample using a linear mixed model approach to account for familial relatedness and population substructure. Results were then meta-analyzed using a weighted inverse-variance fixed-effects model in METAL (Willer et al., 2010) yielding a final sample size of N=1,568. Notably, no individual variants in either of the respective samples, or the combined meta-analysis of the two samples, reached genome-wide significance. The top single-variant association across samples was found in an intergenic region located near the CD5L gene (rs12086663, p= 6.28x10^-7) located on chromosome 1. An additional finding of interest emerged in an intronic region of SYNE1 located on chromosome 6 (rs4472361, p= 2.64x10^-6) that has previously been implicated with alcohol dependence and alcohol and tobacco co-use outcomes (e.g., Edenberg et al., 2010; Otto et al., 2017). These findings suggest that, similar to other complex traits, larger sample sizes will be required to achieve adequate power to examine the genetic influences contributing to LR to alcohol. Importantly, the current study represents initial findings from a larger meta-analytic effort aimed toward achieving sufficient power to identify individual genetic variants that contribute to alcohol response outcomes and AUD more broadly.