GGtools (version 5.8.0)

ciseqByCluster: end-to-end cluster-based cis-eQTL search, and allied utilities

Description

end-to-end cluster-based cis-eQTL search, and allied utilities

Usage

ciseqByCluster(cl, pack = "yri1kgv", outprefix = "yrirun",  finaltag = "partyri100k", chromsToRun = 1:22,  targetfolder = "/freshdata/YRI_3", radius = 100000L, nperm = 3L, ncoresPerNode = 8,  numPCtoFilter = 10, lowerMAF = 0.02, geneannopk = "lumiHumanAll.db",  snpannopk = "SNPlocs.Hsapiens.dbSNP144.GRCh37", smchrpref = "chr", tmpForSort = "/tmp", numtiles = 200, postProcCores = 12, reqlist = NULL)

Arguments

cl

instance of S3 cluster class from parallel package

pack

character string naming package to which getSS can be applied to generate smlSet-class instances

outprefix

character string used to prefix names of output GFF3 files

finaltag

character string used to prefix names of final amalgamated GFF3 and data.table instances

chromsToRun

numeric tags of chromosomes to be analyzed

targetfolder

character string naming folder where GFF3 will be deposited

radius

extent of search around gene model in bp

nperm

number of permutations for plug-in FDR computation (usually a small integer)

ncoresPerNode

number of cores for multicore testing: chromosomes map to nodes, genes map to cores

numPCtoFilter

number of PCs to be removed through clipPCs

lowerMAF

lower bound on MAF of SNP to be included for testing

geneannopk

character string naming Bioconductor package with annotation for expression probe identifiers

snpannopk

character string naming Bioconductor package with annotation for SNP locations

smchrpref

character prefix converting chromsToRun elements to basenames of rda files harboring SnpMatrix instances

tmpForSort

the assembly of final resources employs unix sort, and substantial temporary space can be required; this parameter tells where the temp files will reside

numtiles

number of tiles into which the genome in use will be sliced for parallel processing in final assembly

postProcCores

numeric establishing number of cores to use for final assembly of annotated output

reqlist

rescue request, see Details section

Value

a set of GFF3 files encoding all cis associations with location and various metadata

Details

purpose is to maximize throughput of cis-eQTL testing in a two-level concurrent computing environment, where a cluster as defined in package parallel has nodes to which half-chromosomes will be dispatched; each node is assumed to be multicore and genes are mapped to cores during the iteration process.

the reqlist parameter consists of a list of elements (chromosome name, subchromosome token, and handler) to be used for completing a partial run

See Also

gffprocess, cgff2dt

Examples

#none yet