genbin

genbin (GENetics BINaries) provides well-tested wrappers for narrow functions of several binary packages, including plink1 and plink2 (limited capabilities), gcta, bolt, gemma, and emmax. Focuses on approaches for genetic association, population structure (kinship/GRM matrices and PCA), and heritability estimation. Outputs are omitted unless there are errors (for easier debugging). Parsers for the various output tables are also provided. Complements the genio package (which focuses on parsing and not on binary wrappers).

Installation

The current development version can be installed from the GitHub repository using devtools:

install.packages("devtools") # if needed
library(devtools)
install_github('OchoaLab/genbin')

Examples

Here are two minimal pipelines we can run through R! Several more pipelines are supported but not shown here. The package so far assumes a Linux system, the only case that has been tested; other operating systems may or may not work.

library(genbin)
library(genio)

# examples below assume these "plink" files exist:
# name.bed, name.bim, name.fam, name.phen
name <- 'name'
# number of PCs for some examples:
n_pcs <- 10

GCTA pipeline

The GCTA examples assume that gcta64 is a binary in the system's PATH.

# create GRM from plink files
gcta_grm( name )
# optional: read kinship matrix into R
data <- genio::read_grm( name )
kinship <- data$kinship

# perform mixed linear model association, returning table
data <- gcta_mlma( name )
# association p-values
data$p

# optional: calculate PCs (creates eigenvec/eigenval files)
gcta_pca( name, n_pcs = n_pcs )
# read PCs into R
data <- genio::read_eigenvec( name )
# delete eigenvec/eigenval files
delete_files_pca( name )

# cleanup
# delete GRM files
genio::delete_files_grm( name )
# delete association table
delete_files_gcta_mlma( name )
# delete log
delete_files_log( name )

PLINK PCA pipeline

The plink examples assume that plink2 is a binary in the system's PATH.

# calculate PCs (creates eigenvec/eigenval files)
plink_pca( name, n_pcs = n_pcs )
# optional: read PCs into R
data <- genio::read_eigenvec( name )

# perform PCA association, returning table
data <- plink_glm( name, file_covar = paste0( name, '.eigenvec' ) )
# association p-values
data$p

# cleanup
# delete eigenvec/eigenval files
delete_files_pca( name )
# delete association table
delete_files_plink_glm( name )
# delete log
delete_files_log( name )