GGtools (version 5.8.0)

eqtlTests.me: use MatrixEQTL computations and statistics as a back end to GGtools eqtlTests

Description

use MatrixEQTL computations and statistics as a back end to GGtools eqtlTests

Usage

eqtlTests.me(smlSet, rhs = ~1, runname = "20",  targdir = "cisScratch.me", pvot = 0.5, geneApply = lapply,  shortfac = 100, checkValid = TRUE, useUncertain = TRUE,  glmfamily = "gaussian", scoretx = abs,  matrixEQTL.engine.control =  list(output_file_name = "/dev/null", useModel = modelLINEAR,  errorCovariance = numeric(),  verbose = FALSE,  pvalue.hist = FALSE),  snpSlicedData.control = .slicedDataDefaults,  geneSlicedData.control = .slicedDataDefaults,  covarSlicedData.control = .slicedDataDefaults,  covariates_file_name = character())

Arguments

smlSet

instance of smlSet-class

rhs

formula for adjustment of tests for covariates or stratification, see snp.rhs.tests

runname

tag used to distinguish emitted files

targdir

folder where ff archives will reside

pvot

setting for pvOutputThreshold in Matrix_eQTL_engine

geneApply

lapply-like function for iteration over genes, mclapply is suitable when in multicore environments

shortfac

scaling factor to increase precision when test results are stored as short ints in ff

checkValid

logical to check validity of input smlSet

useUncertain

logical informing snp.rhs.tests that imputed real-valued B allele counts may be present among genotype data

glmfamily

family specification for snp.rhs.tests

scoretx

function to be applied to MatrixEQTL statistics. Defaults to abs, for signless association testing

matrixEQTL.engine.control

list of parameters passed to Matrix_eQTL_engine

snpSlicedData.control

list of parameters used to define SlicedData-class instances

geneSlicedData.control

list of parameters used to define SlicedData-class instances

covarSlicedData.control

list of parameters used to define SlicedData-class instances

covariates_file_name

if covariates are to be used with MatrixEQTL testing engine, they reside in this file. regressOut can be used to avoid this if plug-in FDR are to be used

Value

see eqtlTests

Details

provisional interface

References

Shabalin et al Bioinformatics (OUP) 2012

Examples

if (require(MatrixEQTL)) {
 g22 = nsFilter( chrFilter( getSS("GGdata", "22"), "22" ), var.cutoff = .8 )
 m22 = eqtlTests.me(g22)
}