Streamed Genomic Relationship Matrix — hbr_build_grm • HaploBreedR

Computes the VanRaden (2008) Method 1 genomic relationship matrix in streaming blocks directly from binary PLINK files.

Usage

hbr_build_grm(
  plink_prefix,
  maf_min = 0.05,
  block_cols = 20000L,
  n_cores = 1L,
  verbose = TRUE,
  train_idx = NULL
)

Arguments

plink_prefix: Character. Path prefix to PLINK binary files.
maf_min: Numeric. Minimum minor allele frequency (default 0.05).
block_cols: Integer. SNPs evaluated per memory chunk (default 20,000).
n_cores: Integer. Worker threads for parallel block evaluation.
verbose: Logical. Print progress.
train_idx: Optional integer vector of individuals for allele frequency estimation.

Value

List containing `A`, `d`, `p`, `keep`, `n`, `m`, and `snps`.

Examples

toy <- system.file("extdata", "toy", package = "HaploBreedR")
if (requireNamespace("BEDMatrix", quietly = TRUE)) {
  grm <- hbr_build_grm(file.path(toy, "toy"), verbose = FALSE)
  length(grm$snps)
}
#> [1] 50