PRSice-2 is a clumping and thresholding-based tool used for calculating Polygenic Risk Scores (PRS) across a range of p-value thresholds.
In this project, PRSice-2 was applied to the UK Biobank dataset to:
- Benchmark PRS results alongside
pgsc_calcandPRScs - Generate scores across multiple p-value thresholds
- Requirements: R and PLINK 2.0
Download the executable:
wget https://github.com/choishingwan/PRSice/releases/download/2.3.5/PRSice_linux.zip
unzip PRSice_linux.zip
chmod +x PRSice_linuxThe following packages are needed in R:
install.packages("data.table")
install.packages("ggplot2")
install.packages("pROC")
install.packages("optparse")
install.packages("broom")- Base (GWAS) data: Summary statistics file with SNPs and their effect sizes
- Target data: Genotype data in PLINK Binary or BGEN format
- Phenotype : Individual IDs with their case/control information
- Bar Plot for various p-value thresholds
- High Resolution Plot for various p-value thresholds
- Quantile Plot with increasing PRS
.prsicefile with the PRS Model fit.summaryfile.logfile for each run.bestfile with final PRS for each individual
Rscript PRSice.R \
--base <gwas> \
--target <genotypes> \
--pheno <phenotype file> \
--binary-target <T/F> \
--prsice PRSice_linux \
--out <output_dir>Additional flags can be included depending on the phenotype, datasets and analysis goals.