Visualize gene level behavior of genes within a geneset across a contrast.

It is informative to look at the individual log fold changes of the genes within a gene set to explore the degree to which they (1) are coherent with respect to each other; and (2) see how the compare to the background distribution of log fold changes of the entire transcriptome.

You can visualize this behavior via a type = "density" plot, or a type = "boxplot". It is also common to plot either the individual log fold changes value = "logFC"or t-statisticsvalue = "t"`.

iplot(
  x,
  name,
  value = "logFC",
  type = c("density", "gsea", "boxplot"),
  tools = c("wheel_zoom", "box_select", "reset", "save"),
  main = NULL,
  with.legend = TRUE,
  collection = NULL,
  result_name = NULL,
  stats.report = c("NES", "logFC", "pval", "padj", "n"),
  shiny_source = "mggenes",
  width = NULL,
  height = NULL,
  ggtheme = ggplot2::theme_bw(),
  trim = 0.005,
  interactive = TRUE,
  ...
)

Arguments

x: A SparrowResult() object
name: the name of the geneset to plot
value: A string indicating the column name for the value of the gene-level metadata to plot. Default is "logFC". Anoter often used choice might also be "t", to plot t-statistics (if they're in the result). But this can be any numeric column found in the data.frame returned by geneSet(x, y, j). If this is a named string (vector), then the value in names(value) will be used on the axis when plotted.
type: plot the distributions as a "density" plot or "boxplot".
tools: the tools to display in the rbokeh plot
main: A title to display. If not specified, the gene set name will be used, otherwise you can pass in a custom title, or NULL will disable the title altogether.
with.legend: Draws a legend to map point color to meaning. There are three levels a point (gene level statistic) can be color as, "notsig", "psig", and "sig". "notsig" implies that the FDR >= 10%, "psig" means that FDR <= 10%, but the logFC is "unremarkable" (< 1), and "sig" means that both the FDR <= 10% and the logFC >= 1
collection: If you have genesets with duplicate names in x (only possible with a GeneSetDb object), provide the name of the collection here to disambiguate (default: NULL).
result_name: if not NULL (the default), it fetches statistics for the given GSEA result using result(x, name = result_name, ...) which can then be used for plotting.
shiny_source: the name of this element that is used in shiny callbacks. Defaults to "mggenes".
width, height: the width and height of the output plotly plot
ggtheme: a ggplot theme, like the thing returned from ggplot2::theme_bw(), for instance.
trim: used to define the upper and lower quantiles to max out the individual gene statistics in the selected geneset.
...: pass through parameters to internal boxplot/density/gsea plotting functions

Value

the ploty plot object

Examples

mgr <- exampleSparrowResult()
upreg <- "SOTIRIOU_BREAST_CANCER_GRADE_1_VS_3_UP"
dnreg <- "TURASHVILI_BREAST_LOBULAR_CARCINOMA_VS_DUCTAL_NORMAL_DN"

iplot(mgr, upreg, value = c("t-statistic" = "t"), type = "density")

iplot(mgr, upreg, value = c("t-statistic" = "t"), type = "density", .plot_static = TRUE)
#> Error in iplot.density.plotly(x, dat, value, main, with.legend = with.legend,     tools = tools, shiny_source = shiny_source, ggtheme = ggtheme,     trim = trim, result_name = result_name, gstats = gstats,     .plot_static = !interactive, ...): formal argument ".plot_static" matched by multiple actual arguments
iplot(mgr, upreg, value = c("log2FC" = "logFC"), type = "boxplot")

iplot(mgr, upreg, value = c("t-statistic" = "t"), type = "gsea")
#> Loading required namespace: testthat
#> Loading required namespace: fgsea

iplot(mgr, dnreg, value = c("t-statistic" = "t"), type = "gsea")