geneAttribution: geneAttribution: Identification of candidate genes associated...

In geneAttribution: Identification of candidate genes associated with genetic variation

Description

Identification of the most likely gene or genes through which variation at a given genomic locus in the human genome acts. The most basic functionality assumes that the closer gene is to the input locus, the more likely the gene is to be causative. Additionally, any empirical data that links genomic regions to genes (e.g. eQTL or genome conformation data) can be used if it is supplied in UCSC .bed file format. A typical workflow requires loading gene models and empirical data, then running geneAttribution() on the locus of interest

Given genomic coordinate, return normalized probability for each gene

Usage

geneAttribution(chr, pos, geneCoordinates, empiricalData, lambda = 7.61e-06,
  maxDist = 1e+06, minPP = 0.01)

Arguments

chr	A character string representing a chromosome (e.g. "chr2")
pos	An integer representing a genomic position in the same genome build that gene models
geneCoordinates	A GenomicRanges object, as generated by geneModels()
empiricalData	A list of GenomicRanges objects, as generated by loadBed(). Optional
lambda	Float. Variable for exponential distribution. Default based on empirical eQTL data from multiple tissues. Optional
maxDist	Integer. Only genes within this distance of query locus are considered. Optional
minPP	Float. Genes with a posterior probability < minPP are lumped as "Other". Can be set to 0 when all genes should be reported. Optional

Value

A sorted, numeric vector of normalized gene probabilities

Examples

geneLocs <- geneModels()
fileName <- system.file("extdata", "eqtlHaplotypeBlocks.b38.bed", package="geneAttribution")
empirical <- loadBed(fileName)
geneAttribution("chr2", 127156000, geneLocs, empirical)

geneAttribution documentation built on Nov. 8, 2020, 5:17 p.m.