catplot.GSEA.random: Compare multiple GSEA runs via CAT plots & estimate the...

In drmjc/metaGSEA: Meta-analysis of GSEA analyses, including intra- and inter-experiment comparisons

Description

This function generates a normal X vs Y CAT plot, but then randomises Y, B times to get a permuted distribution of overlaps.

Usage

catplot.GSEA.random(..., labels = LETTERS, B = 100, sizes = NULL,
  sizes.random = sizes, random.col = "grey", ylim = c(0, 1),
  proportion = TRUE, pch = 1, legend.pos = "bottomright")

Arguments

labels	the name of of each GSEA run. must be same length as ..., or gsea.list.
B	the number of permutations. default=100
sizes	Controls the granularity of the CAT plot. see details
sizes.random	Controls the granulatity of the randomised CAT plot. Since this gets done B times, this setting can really impact the speed of this function. See details.
random.col	the fill colour for the random background
ylim	see par
proportion	if TRUE, x-axis is proportion of all genesets, if FALSE (default), the number of genesets is plotted.
pch	the print character. see par
legend.pos	the legend position. see legend
...	at least 2 GSEA objects

Details

Controlling the granularity of the CAT plot with sizes and sizes.random
“sizes” controls the granularity of the CAT plot. If sizes=NULL, then sizes = 1,2,3,4,...,N, where N is the total number of genesets. This gives a very detailed CAT plot, but if this is too slow, try setting sizes=seq(0,N,5). Same goes for sizes.random, but since there are B randomisations, this setting has more of an impact upon the speed of this code. typically with just ~1400 genesets, it's fast enough to calculate on all values from 1 to N, however: sizes=seq(0,N,5), or sizes=seq(0,N,10) will be much quicker.

Author(s)

Mark Cowley

See Also

catplot.vs.random, catplot.GSEA

drmjc/metaGSEA documentation built on Aug. 8, 2020, 1:53 p.m.