Next we apply PRSice to two tobacco-related phenotypes from the TAG consortium (Thorgeirsson et al., 2013) and the RADIANT-UK MDD data. These analyses reveal, for the first time, shared genetic aetiology between the dichotomous trait ‘ever smoked’ and MDD, but not between smoking consumption, as a quantitative trait, and MDD (Supplementary Fig. S1). In the former, high-resolution scoring again produces a substantially different best-fit PRS than that from broad PT, in terms of model fit, significance and number of SNPs included (Supplementary Fig. S1b).
