Plots log-Fold Change versus log-Concentration (or, M versus A) for Count Data
Both of these functions plot the log-fold change (i.e. the log of the ratio of expression levels for each gene between two experimential groups) against the log-concentration (i.e. the overall average expression level for each gene across the two groups). To represent counts that were low (e.g. zero in 1 library and non-zero in the other) in one of the two conditions, a 'smear' of points at low A value is presented in
plotSmear(object, pair=NULL, de.tags=NULL, xlab="Average logCPM", ylab="logFC", pch=19, cex=0.2, smearWidth=0.5, panel.first=grid(), smooth.scatter=FALSE, lowess=FALSE, ...)
DGELRT object containing data to produce an MA-plot.
pair of experimental conditions to plot (if
NULL, the first two conditions are used). Ignored if
object is a
rownames for genes identified as being differentially expressed; use
glmLRT to identify DE genes. Note that `tag' and `gene' are synonymous here.
scalar or vector giving the character(s) to be used in the plot; default value of
19 gives a round point.
character expansion factor, numerical value giving the amount by which plotting text and symbols should be magnified relative to the default; default
cex=0.2 to make the plotted points smaller
width of the smear
an expression to be evaluated after the plot axes are set up but before any plotting takes place; the default
grid() draws a background grid to aid interpretation of the plot
logical, whether to produce a 'smooth scatter' plot using the
KernSmooth::smoothScatter function or just a regular scatter plot; default is
FALSE, i.e. produce a regular scatter plot
logical, indicating whether or not to add a lowess curve to the MA-plot to give an indication of any trend in the log-fold change with log-concentration
further arguments passed on to
plotSmear is a more sophisticated and superior way to produce an 'MA plot'.
plotSmear resolves the problem of plotting genes that have a total count of zero for one of the groups by adding the 'smear' of points at low A value. The points to be smeared are identified as being equal to the minimum estimated concentration in one of the two groups. The smear is created by using random uniform numbers of width
smearWidth to the left of the minimum A.
plotSmear also allows easy highlighting of differentially expressed (DE) genes.
y <- matrix(rnbinom(10000,mu=5,size=2),ncol=4)
d <- DGEList(counts=y, group=rep(1:2,each=2), lib.size=colSums(y))
rownames(d$counts) <- paste("gene",1:nrow(d$counts),sep=".")
d <- estimateCommonDisp(d)
# find differential expression
de <- exactTest(d)
# highlighting the top 500 most DE genes
de.genes <- rownames(topTags(de, n=500)$table)