Plot where a Venn diagram misrepresents the data

Redraws the Venn diagram produced by plotVenn, shaded by the signed error of each region, so that the regions the diagram gets wrong can be read off the picture rather than inferred from a single summary statistic. This is a thin wrapper around error_plot.

Usage

plotVennError(venn, ...)

Arguments

venn

A plot returned by plotVenn, carrying the "euler_fit" attribute that function attaches. Note that this is the plotVenn() output, not the GenomicOverlapResult or SetOverlapResult that plotVenn() itself takes.

...

Additional arguments passed to error_plot.

Value

The plot returned by error_plot, a gTree that can be drawn, arranged with other grobs, or written out with saveViz.

Details

Where diagError reports how far off the worst region is, this plot shows which regions are off and in which direction, with each region's regionError printed inside it. Use it when diagError exceeds 1e-6.

See also

plotVenn for the diagram itself and its fit diagnostics, plotUpSet for a representation that does not approximate any count by an area.

Examples

data(gene_list)
res_sets <- computeOverlaps(gene_list)

# Where does the diagram misrepresent the counts?
venn <- plotVenn(res_sets)
#> ✔ Venn diagError = 1.728e-14  (<= 1e-06)
#>   Access fit diagnostics with attr(<plotVenn output>, "fit_diagnostics")
plotVennError(venn)