glMDPlot.MArrayLM: Glimma MD Plot
In Glimma: Interactive HTML graphics

Description Usage Arguments Value Author(s) Examples

Draw an interactive MD plot from a MArrayLM object

## S3 method for class 'MArrayLM'
glMDPlot(
  x,
  counts = NULL,
  anno = NULL,
  groups = NULL,
  samples = NULL,
  status = rep(0, nrow(x)),
  transform = FALSE,
  main = "",
  xlab = "Average log CPM",
  ylab = "log-fold-change",
  side.main = "GeneID",
  side.xlab = "Group",
  side.ylab = "Expression",
  side.log = FALSE,
  side.gridstep = ifelse(!transform || side.log, FALSE, 0.5),
  coef = ncol(x$coefficients),
  p.adj.method = "BH",
  jitter = 30,
  display.columns = NULL,
  cols = c("#00bfff", "#858585", "#ff3030"),
  sample.cols = rep("#1f77b4", ncol(counts)),
  path = getwd(),
  folder = "glimma-plots",
  html = "MD-Plot",
  launch = TRUE,
  ...
)

`x`	the MArrayLM object.
`counts`	the matrix of expression values, with samples in columns.
`anno`	the data.frame containing gene annotations.
`groups`	the factor containing experimental groups of the samples.
`samples`	the names of the samples.
`status`	vector giving the control status of data point, of same length as the number of rows of object. If NULL, then all points are plotted in the default colour.
`transform`	TRUE if counts should be log-cpm transformed.
`main`	the title for the left plot.
`xlab`	label for x axis on left plot.
`ylab`	label for y axis on left plot.
`side.main`	the column containing mains for right plot.
`side.xlab`	label for x axis on right plot.
`side.ylab`	label for y axis on right plot.
`side.log`	TRUE to plot expression on the right plot on log scale.
`side.gridstep`	intervals along which to place grid lines on y axis. Currently only available for linear scale.
`coef`	integer or character index vector indicating which column of object to plot.
`p.adj.method`	character vector indicating multiple testing correction method. See `p.adjust` for available methods. (defaults to "BH")
`jitter`	the amount of jitter to apply to the samples in the expressions plot.
`display.columns`	character vector containing names of columns to display in mouseover tooltips and table.
`cols`	vector of strings denoting colours corresponding to control status -1, 0 and 1. (may be R named colours or Hex values)
`sample.cols`	vector of strings denoting colours for each sample point on the expression plot.
`path`	the path in which the folder will be created.
`folder`	the name of the fold to save html file to.
`html`	the name of the html file to save plots to.
`launch`	TRUE to launch plot after call.
`...`	additional arguments to be passed onto the MD plot. (main, xlab, ylab can be set for the left plot)

Draws a two-panel interactive MD plot in an html page. The left plot shows the log-fold-change vs average expression. The right plot shows the expression levels of a particular gene of each sample. Hovering over points on left plot will plot expression level for corresponding gene, clicking on points will fix the expression plot to gene. Clicking on rows on the table has the same effect as clicking on the corresponding gene in the plot.

Shian Su

library(limma)
library(edgeR)

data(lymphomaRNAseq)
x <- lymphomaRNAseq

sel <- rowSums(cpm(x$counts)>0.5)>=3
x <- x[sel,]

genotype <- relevel(x$samples$group, "Smchd1-null")
x <- calcNormFactors(x, method="TMM")
des <- model.matrix(~genotype)

## Apply voom with sample quality weights and fit linear model
v <- voomWithQualityWeights(x, design=des, plot=FALSE)
vfit <- lmFit(v,des)

## Apply treat relative to a fold-change of 1.5
vtfit <- treat(vfit,lfc=log2(1.5))
vfit <- eBayes(vfit)
results <- decideTests(vfit,p.value=0.01)


glMDPlot(vfit, counts=x$counts, anno=x$genes, groups=genotype, samples=1:7,
         status=results[,2], main="MD plot: Wild-type vs Smchd1",
         display.columns=c("Symbols", "GeneID", "GeneName"),
         folder="Smchd1-Lymphoma")