/Rprs-1

R package for computing Polygenic Risk Scores (PRS)

Primary LanguageRGNU General Public License v3.0GPL-3.0

Rprs

R package for computing Polygenic Risk Scores (PRS) from dosages (DS). Supports VCF and BCF input formats.

Install

Clone repository using git clone and run the following command:

R CMD INSTALL Rprs

Run

P-value Thresholding (PT) method.

  • Uses an LD clumped weight file as input
  • Weights file must have the following columns (in the same order): CHROM, POS, OA, EA, WEIGHT, PVALUE.
  • If no samples file is provided (i.e. NULL), then all samples are used.
library(Rprs)
p <- prs_pt("weights.txt", "dosages.vcf.gz", "samples.txt", pvalues = c(1.0, 5e-8))
write.table(p, col.names = TRUE, row.names = FALSE, sep = "\t", quote = F)

Genome-wide PRS method.

  • Weights file must have the following columns (in arbitrary order): CHROM, POS, OA, EA, and multiple columns for weights.
  • All weight columns must have the same prefix (e.g. "LDpred_", "LDpred", "WEIGHT_", "WEIGHT", or similar). You must specify this prefix using the weight_col function argument. Default is WEIGHT_.
  • If no samples file is provided (i.e. NULL), then all samples are used.
library(Rprs)
p <- prs_gw("weights.txt", "dosages.vcf.gz", "samples.txt", weight_col = "LDpred_")
write.table(p, col.names = TRUE, row.names = FALSE, sep = "\t", quote = F)