PWMEnrich (version 4.8.2)

makeBackground: Make a background for a set of position frequency matrices

Description

This is a convenience front-end function to compile new backgrounds for a set of PFMs. Currently only supports D. melanogaster, but in the future should support other common organisms as well.

Usage

makeBackground(motifs, organism = "dm3", type = "logn", quick = FALSE, bg.seq=NULL, ...)

Arguments

motifs

a list of position frequency matrices (4xL matrices)

organism

either a name of the organisms for which the background should be compiled (currently supported names are "dm3", "mm9" and "hg19"), or a BSgenome object (see BSgenome package).

type

the type of background to be compiled. Possible types are:

• "logn" - estimate a lognormal background
• "cutoff" - estimate a Z-score background with fixed log-odds cutoff (in log2)
• "pval" - estimate a Z-score background with a fixed P-value cutoff. Note that this may require a lot of memory since the P-value of motif hits is first estimated from the empirical distribution.
• "empirical" - create an empirical P-value background. Note that this may require a lot of memory (up to 10GB in default "slow" mode (quick=FALSE) for 126 JASPAR motifs and 1000 D. melanogaster promoters).
• "GEV" - estimate a generalized extreme value (GEV) distribution background by fitting linear regression to distribution parameters in log space

quick

if to preform fitting on a reduced set of 100 promoters. This will not give as good results but is much quicker than fitting to all the promoters (~10k). Usage of this parameter is recommended only for testing and rough estimates.

bg.seq

a set of background sequences to use. This parameter overrides the "organism" and "quick" parameters.

...

other named parameters that backend function makePWM***Background functions take.

Examples

# load in the two example de-novo motifs
motifs = readMotifs(system.file(package="PWMEnrich", dir="extdata", file="example.transfac"), remove.acc=TRUE)

## Not run: 
#   # construct lognormal background
#   bg.logn = makeBackground(motifs, organism="dm3", type="logn")
# 
#   # alternatively, any BSgenome object can also be used
#   if(require("BSgenome.Dmelanogaster.UCSC.dm3"))
#     bg.logn = makeBackground(motifs, organism=Dmelanogaster, type="logn")
# 
#   # construct a Z-score of hits with P-value background
#   bg.pval = makeBackground(motifs, organism="dm3", type="pval", p.value=1e-3)
# 
#   # now we can use them to scan for enrichment in sequences (in this case there is a consensus Tin binding site)
#   motifEnrichment(DNAString("TGCATCAAGTGTGTAGTG"), bg.logn)
#   motifEnrichment(DNAString("TGCATCAAGTGTGTAGTG"), bg.pval)
# ## End(Not run)