findMotif: De-novo discovery of distriminative motifs

In motifRG: A package for discriminative motif discovery, designed for high throughput sequencing dataset

Description

The function searches motifs that discriminate the given foregound and background sequences.

Usage

findMotif(all.seq,  category, weights = rep(1, length(all.seq)),
start.width=6,min.cutoff=5, min.ratio=1.3,
min.frac=0.01, both.strand=TRUE, flank=2, max.motif=5,
mask=TRUE,other.data=NULL, start.nmer=NULL,
enriched.only=F,n.bootstrap = 5, bootstrap.pvalue=0.1,is.parallel =
TRUE,mc.cores = 4,min.info=10,max.width=15,discretize=TRUE)

Arguments

all.seq	DNAStringSet; foreground and background sequences.
category	numeric vector; specify which sequences are foreground (with value 1), and background (value 0).
weights	numeric vector: the weights for all sequences. Default: 1
start.width	logical; the width for enumerating seed patterns
min.cutoff	numeric; the score cutoff required for seed selection. All scores are negative, the lower the better.
min.ratio	numeric; the minimum fold change of motif occurences in foreground vs background.
min.frac	numeric; the minimum fraction of fg/bg sequences containing the candidate motifs
both.strand	logical; if true, search both strands
flank	integer; the length for step-wise pattern extension at both ends on candidate motifs
max.motif	integer; the maximum number of output motifs
mask	logical; if true, mask previous motifs when searching for the next motif
other.data	if not NULL, a matrix with additional terms for the regression model for bias adjustment
start.nmer	if not NULL, a matrix with counts for user specified seed pattern in each sequence
enriched.only	logical; if true, only predict enriched motif
n.bootstrap	integer; the number of bootstrapping tests to estimate score variance
bootstrap.pvalue	numeric: the bootstrap t.test pvalues to determine the significance of improvement
is.parallel	logical;if true, runs in parallel mode, and requires "parallel" library
mc.cores	integer; the number of CPUs for paralel run
min.info	minimal information content for the motif to prevent it from being too degenerate
max.width	maximum width of the motif for extension
discretize	logical default TRUE

Value

return a list with following elements:

motifs

a list motif descriptions of class Motif-class

.

category	input binary specification of foreground/background
mask.motifs	if mask=T, then mask.motifs contain the description of motif is based on motif matches after the input sequences being masked by previous motifs. In this case, "motifs" contained the unmasked motif descriptions.

Examples

MD.peak.seq <- readDNAStringSet(system.file("extdata","MD.peak.fa", package="motifRG"))
MD.control.seq <- readDNAStringSet(system.file("extdata","MD.control.fa", package="motifRG"))
category <- c(rep(1, length(MD.peak.seq)), rep(0, length(MD.control.seq)))
MD.motifs <- findMotif(append(MD.peak.seq, MD.control.seq),category, max.motif=3,enriched=TRUE)

### Get summary of motifs
summaryMotif(MD.motifs$motifs, MD.motifs$category)

### plot the dinucleotide representation of the first motif
plotMotif(MD.motifs$motifs[[1]]@match$pattern)

### Create table of motifs in Latex 
motifLatexTable(MD.motifs, main="MD motifs")

### Create table of motifs in Html
motifHtmlTable(MD.motifs)

motifRG documentation built on April 28, 2020, 8:46 p.m.