Genetic Heterogeneity Across Dimensions of Alcohol Use Behaviors

Abstract

AbstractBackgroundIncreasingly large samples in genome-wide association studies (GWAS) for alcohol use behaviors (AUBs) have led to an influx of implicated genes, yet the clinical and functional understanding of these associations remains low. This is, in part, because most GWASs do not account for complex and varied manifestations of AUBs. This study applied a multidimensional framework to investigate the latent genetic structure underlying heterogeneous dimensions of AUBs.MethodsMulti-modal assessments (self-report, interview, electronic health records) were obtained from approximately 400,000 UK Biobank participants. GWAS was conducted for 18 distinct AUBs, including consumption, drinking patterns, alcohol problems, and clinical sequelae. Latent genetic factors were identified and carried forward to GWAS using genomic structural equation modeling, followed by functional annotation, genetic correlation, and enrichment analyses to interpret the genetic associations.ResultsFour latent factors were identified:Problems, Consumption, BeerPref(declining alcohol consumption with a preference for drinking beer), andAtypicalPref(drinking fortified wine and spirits). The latent factors were moderately correlated (rg=.12-.57) and had distinct patterns of associations, withBeerPrefin particular implicating many novel genomic regions. Patterns of regional and cell type specific gene expression in the brain also differed between the latent factors.ConclusionDeep phenotyping and multi-modal assessment is an important next step to improve understanding of the genetic etiology of AUBs, in addition to increasing sample size. Further effort is required to uncover the genetic heterogeneity underlying AUBs using methods that account for their complex, multidimensional nature.