qgg (version 1.0.3)

grm: 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 = 0, 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=0 is default) used for missing values in the genotype data

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

Examples

# NOT RUN {
# 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)

# }
# NOT RUN {
# Eigen value decompostion GRM
eig <- grm(GRM=GRM, task="eigen")

# }