transAssoc: compute 'trans' SNP-feature associations by wrapping AllAssoc

In gQTLstats: gQTLstats: computationally efficient analysis for eQTL and allied studies

Description

compute 'trans' SNP-feature associations by wrapping AllAssoc, retaining only the strongest associations (and similarly filtered association scores computed under permutation)

Usage

transAssoc(variantGR, exSE, vcfgen, bufsize = 10, nperm = 3, exChLen = 2 * bufsize, ...)

Arguments

variantGR	GRanges instance establishing scope of variants to test
exSE	SummarizedExperiment instance, all of whose features will be tested for association with all SNP
vcfgen	a function returning a path to a tabix-indexed VCF file from which SNP genotypes will be extracted
bufsize	Size of 'buffer' used to retain largest feature association scores encountered during the search. The scores and the names of associated genes are retained in 'scorebuf' and 'elnames' components of output GRanges
nperm	number of permutations of features against genotypes to be performed for realizing null distribution of association scores
exChLen	size of chunks of exSE to be tested through calls to AllAssoc; this is intended to allow control of RAM usage
...	arguments passed to AllAssoc

Value

a GRanges with mcols including

Examples

## Not run:   # requires acccess to 1KG S3
library(geuvPack)
data(geuFPKM)
seqlevelsStyle(geuFPKM) = "NCBI"
mysr = GRanges("20", IRanges(33000055, 33020055))
genome(mysr) = "hg19"
tt = transAssoc(mysr, geuFPKM[1:16,],
    bufsize=3, exChLen=4, vcfgen=function(x)gtpath(paste0("chr", x)) )
colnames(mcols(tt))
table(as.character(mcols(tt)$elnames))

## End(Not run)

gQTLstats documentation built on Nov. 8, 2020, 7:53 p.m.