estimated that the total variance explained by common SNPs for adult alcohol use problems was 33%, suggesting that a large portion of the heritability can be explained by variations in common SNPs in aggregate, rather than in isolation (Mbarek et al., 2015). These findings reinforce the idea that complex traits are largely influenced by polygenic variation, rather than one or a few SNPs (Plomin et al., 2009; Purcell et al., 2009). Researchers can index the composite effects of common SNPs by computing polygenic risk scores (PRS), which represent a quantitative distribution of aggregate genetic liability to predict the phenotypic variance in one sample using genome-wide information from a separate, independent sample. The PRS approach was first applied to predict the risk of bipolar disorder and schizophrenia (Purcell et al., 2009) but has since been applied to predict risk for alcohol-related phenotypes (Kos et al., 2013; Salvatore et al., 2014a; Vink et al., 2014), explaining about .63–1.1% of variation in alcohol use behaviors. Although this is a relatively small amount of the total variance, it still constitutes a larger effect than single SNPs confer independently (Plomin et al., 2009).
