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

In PWMEnrich: PWM enrichment analysis

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.

Author(s)

Robert Stojnic, Diego Diez

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(requireNamespace("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)

PWMEnrich documentation built on Nov. 8, 2020, 7:45 p.m.