From 'PLINK' to 'HIrisPlex'

Read 'PLINK' 1.9 binary datasets (BED/BIM/FAM) and generate the CSV files required by the Erasmus MC 'HIrisPlex' / 'HIrisPlex-S' webtool < https://hirisplex.erasmusmc.nl/>. It maps 'PLINK' alleles to the webtool's required 'rsID_Allele' columns (0/1/2/NA). No external tools (e.g., 'PLINK CLI') are required.

hirisplexr

Build HIrisPlex / HIrisPlex-S CSV files directly from PLINK 1.9 BED/BIM/FAM.

Installation (from source)

# in R
install.packages("BEDMatrix")
install.packages("data.table")

# build & install this package (using R CMD build/check/INSTALL or devtools)

Usage

library(hirisplexr)

# Example prefix chosen for documentation purposes
# (provide your own PLINK files at this path)
prefix <- "inst/extdata/testprefix"

write_hirisplex_csv(prefix, panel = "hirisplexs", out = "hirisplexs.csv")

How it works

Genotypes are read on-demand from .bed using [BEDMatrix], which returns dosages of A1 (first allele in .bim).
The HIrisPlex(-S) webtool requires a count (0/1/2) of a specific input allele for each SNP. We map A1/A2 to that input allele; for strand issues we can also consider complements.

Panels

Panel definitions (rsID and required input allele, in the precise order expected by the webtool) are packaged in inst/extdata/hirisplex_panels.csv. They are sourced from the official webtool and manual (see citations below).

Citations

HIrisPlex-S webtool: https://hirisplex.erasmusmc.nl/ (columns and upload format)
HIrisPlex-S user manual (2018), Erasmus MC (instructions and caveats)
- Notes include: count of 0/1/2 per SNP; use NA if missing; rs6119471 may require strand conversion from NCBI forward.