SFfigure: Plot expression as curves

In pkimes/SigFuge: SigFuge

Description

Function for producing various figures corresponding to the SigFuge functional data approach to studying RNA-seq data as expression curves along base positions. The primary input for the function is a read count matrix and GRanges. The default behavior is to identify clusters based on applying SFlabels to a normalized version of the data produced by SFnormalize. If specified, the function will compute a p-value for the significance of the labels by calling the SFpval function.

Usage

  SFfigure(data, locusname, annot = c(), flip.fig = 1,
           label.exon = 1, print.n = 1, data.labels = 0,
           label.colors = c(), flag = 1, lplots = 2,
           log10 = 1, summary.type = "median",
           savestr = c(), titlestr = c(), pval = 1)

Arguments

data	a d x n matrix or data.frame of read counts at d base positions for n samples.
locusname	a character string specifying gene or locus name to be used in figure title.
annot	a GRanges object or data.frame including annotation information for locus, including: start start of contiguous genomic regions end end of contiguous genomic regions seqname chromosome name for genomic region strand strandedness of sequence
flip.fig	an indicator whether to flip the plotting direction of the locus if strand == "-" when annotation information is provided.
label.exon	an indicator whether to print the exon boundaries to the figure.
print.n	an indicator whether to print cluster sizes.
data.labels	a n x 1 vector of class labels to use instead of calcuating SigFuge labels
label.colors	a K x 3 matrix of RBG colors specifying cluster colors for K clusters. ggplot2 default colors are used if not specified. If using SigFuge default labels, K=3 even if no low expression samples are flagged.
flag	a n x 1 logical vector of samples flagged as low expression. If flag == 1, default low expression cutoffs are used. If flag == 0, no samples are flagged as low expression (equivalent to setting flag = rep(0, n)).
lplots	a specification of which figures to output 1: curves in single panel, random colors 2: curves in single panel, colored by cluster 3: curves in K panels, separated and colored by cluster 4: curves in n panels, colored by cluster (single sample per panel) 5: cluster medians in single panel, colored by cluster
log10	an indicator whether the y-axis (read depth) should be log10 transformed. Default is to plot on log-scale.
summary.type	a character string specifying which summary statistic should be used when plotting clusters in lplots == 2, 3, and 5. Options: "median" (default) or "mean".
savestr	a string specifying the file name for resulting figures. Extensions can also be specified in savestr. If no extension is specified figures will be saved as pdfs. If length(lplots) > 1, figures will be saved as paste0(savestr,"_x") for x in lplots with the appropirate extension. If no savestr is specified, function will return a list containing the created ggplot objects.
titlestr	a string specifying figure title. If unspecified, default is titlestr=paste(locusname," locus, SigFuge analysis").
pval	an indicator whether the SFpval should be computed. If pval == 1, the p-value is added to the title, i.e. (titlestr=paste0(titlestr, ", p-value = ", p)).

Value

SFfigure returns a figure that is saved to the current working directory if a savestr is specified. Else, a list containing the plots is returned.

Author(s)

Patrick Kimes <[e-mail]>

Examples

# load data
data(geneAnnot)
data(geneDepth)

# only use first 50 samples
mdata <- geneDepth[,1:50]

# make plot
locusname <- "CDKN2A"
SFfigure(mdata, locusname, geneAnnot, flag=1,
 lplots=3, savestr=paste0(locusname,".pdf"), titlestr="CDKN2A locus, LUSC samples",
 pval=1)

mySFs <- SFfigure(mdata, locusname, geneAnnot, flag=1,
            lplots=1, savestr=c(), titlestr="CDKN2A locus, LUSC samples not saved",
            pval=0)
mySFs$plot1

pkimes/SigFuge documentation built on May 25, 2019, 8:19 a.m.