grm: Computing the genomic relationship matrix (GRM)

In psoerensen/qgg: Statistical Tools for Quantitative Genetic Analyses

Computing the genomic relationship matrix (GRM)

Description

The grm function is used to compute a genomic relationship matrix (GRM) based on all, or a subset of marker genotypes. GRM for additive, and non-additive (dominance and epistasis) genetic models can be constructed. The output of the grm function can either be a within-memory GRM object (n x n matrix), or a GRM-list which is a list structure that contains information about the GRM stored in a binary file on the disk.

Usage

grm(
  Glist = NULL,
  GRMlist = NULL,
  ids = NULL,
  rsids = NULL,
  rws = NULL,
  cls = NULL,
  W = NULL,
  method = "add",
  scale = TRUE,
  msize = 100,
  ncores = 1,
  fnG = NULL,
  overwrite = FALSE,
  returnGRM = FALSE,
  miss = NA,
  impute = TRUE,
  pedigree = NULL,
  task = "grm"
)

Arguments

Glist	list providing information about genotypes stored on disk
GRMlist	list providing information about GRM matrix stored in binary files on disk
ids	vector of individuals used for computing GRM
rsids	vector marker rsids used for computing GRM
rws	rows in genotype matrix used for computing GRM
cls	columns in genotype matrix used for computing GRM
W	matrix of centered and scaled genotypes
method	indicator of method used for computing GRM: additive (add, default), dominance (dom) or epistasis (epi-pairs or epi-hadamard (all genotype markers))
scale	logical if TRUE the genotypes in Glist has been scaled to mean zero and variance one
msize	number of genotype markers used for batch processing
ncores	number of cores used to compute the GRM
fnG	name of the binary file used for storing the GRM on disk
overwrite	logical if TRUE the binary file fnG will be overwritten
returnGRM	logical if TRUE function returns the GRM matrix to the R environment
miss	the missing code (miss=NA is default) used for missing values in the genotype data
impute	if missing values in the genotype matrix W then mean impute
pedigree	is a dataframe with pedigree information
task	either computation of GRM (task="grm" which is default) or eigenvalue decomposition of GRM (task="eigen")

Value

Returns a genomic relationship matrix (GRM) if returnGRM=TRUE else a list structure (GRMlist) with information about the GRM stored on disk

Author(s)

Peter Soerensen

Examples

# Simulate data
W <- matrix(rnorm(1000000), ncol = 1000)
	colnames(W) <- as.character(1:ncol(W))
	rownames(W) <- as.character(1:nrow(W))

# Compute GRM
GRM <- grm(W = W)



# Eigen value decompostion GRM
eig <- grm(GRM=GRM, task="eigen")

psoerensen/qgg documentation built on March 9, 2024, 10:02 p.m.