PRSice has been developed to fully automate PRS analyses, substantially expanding the capability of PLINK-1.9 (Chang et al., 2014). A key feature of PRSice is that it can calculate PRS at any number of P-value thresholds (PT) and can thus identify the most predictive (precise) threshold. It requires only GWAS results on a base phenotype and genotype data on a target phenotype as input (base and target phenotype may be the same); it outputs PRS for each individual and figures depicting the PRS model fit at a range of PT. PRSice allows users to include or remove SNPs in linkage disequilibrium, handles genotyped and imputed data, and can calculate ancestry-informative dimensions for use as covariates. These features integrate R code with computations performed in PLINK-1.9 and extensive bash scripts to minimize computational time. PRSice is a command-line program that allows users to apply a large number of PRS, under different parameter settings or across multiple base and target traits. In addition to the standard approach, there is an option to use summary statistics for the target as well as the base data, using the gtx package (Johnson, 2013). For future updates of PRSice, see the website: http://PRSice.info.
