Visualize GSEA results

Usage

VizGSEA(
  gsea_results,
  plot_type = c("summary", "gsea", "heatmap", "dot"),
  gene_ranks = "@gene_ranks",
  gene_sets = "@gene_sets",
  gs = NULL,
  group_by = NULL,
  values_by = "NES",
  signif_by = "p.adjust",
  signif_cutoff = 0.05,
  signif_only = TRUE,
  ...
)

Arguments

gsea_results

A data frame with the results of the fgsea analysis

plot_type

The type of plot to create One of "summary", "gsea"

gene_ranks

A named numeric vector of gene-level rank statistics, with gene identifiers as names. Used to construct the per-term line plots. If a character string starting with "@", the attribute of data with that name (minus the "@") is used as the gene ranks vector.

gene_sets

A named list of gene sets. Each name must correspond to an ID in data, and each element is a character vector of gene identifiers. A GSEA ridge plot is generated for each gene set in the list. If you only want to plot a subset of gene sets, subset the list before passing it to this function. If a character string starting with "@", the attribute of data with that name (minus the "@") is used.

gs

Character vector of gene set IDs to plot. If NULL (default), all gene sets in gene_sets that appear in data$ID are plotted.

group_by

The column name to group by for heatmap and dot plot. They will be used as the columns in the heatmap or dot plot.

values_by

The column name to use for the values in the heatmap or dot plot. Default is "NES" (normalized enrichment score).

signif_by

The column name to use for significance in the heatmap or dot plot. Default is "p.adjust" (adjusted p-value). It can also be "pvalue". If NULL, no significance labels will be added to the heatmap.

signif_cutoff

A numeric vector of significance cutoffs for the heatmap labels. Multiple values can be provided to indicate different levels (at most 3) of significance. For example, c(0.05, 0.01, 0.001) will label pathways with p-values less than 0.05 with "", less than 0.01 with "", and less than 0.001 with "".

signif_only

If TRUE, only pathways that are significant in any group will be kept. The significance is determined by the signif_by column and max(signif_cutoff). If FALSE, all pathways will be kept, but the significance labels will still be added to the heatmap.

...

Additional arguments passed to the plotting function

When plot_type is "summary", they are passed to plotthis::GSEASummaryPlot()
When plot_type is "gsea", they are passed to plotthis::GSEAPlot()

Value

A ggplot object or a list of ggplot objects

Examples

# \donttest{
set.seed(123)
exprs <- matrix(rnorm(1000, 0, 1), nrow = 100, ncol = 10)
colnames(exprs) <- paste0("Sample", 1:10)
rownames(exprs) <- paste0("Gene", 1:100)
classes <- sample(c("A", "B"), 10, replace = TRUE)
ranks <- RunGSEAPreRank(exprs, case = "A", control = "B", classes = classes)
genesets <- list(
    set1 = c("Gene73", "Gene30", "Gene97"),
    set2 = c("Gene4", "Gene5", "Gene6"),
    set3 = c("Gene7", "Gene8", "Gene9"),
    set4 = c("Gene10", "Gene11", "Gene12"),
    set5 = c("Gene13", "Gene14", "Gene15"),
    set6 = c("Gene16", "Gene17", "Gene18"),
    set7 = c("Gene19", "Gene20", "Gene21"),
    set8 = c("Gene22", "Gene23", "Gene24"),
    set9 = c("Gene25", "Gene26", "Gene27"),
    set10 = c("Gene12", "Gene86", "Gene87", "Gene83", "Gene71")
)
r <- RunGSEA(ranks, genesets)

# Visualize the GSEA results
VizGSEA(r, plot_type = "summary")

VizGSEA(r, plot_type = "gsea", gs = c("set10", "set2"))


r$Group <- "A"
r2 <- r
r2$Group <- "B"
r2$NES <- sample(r2$NES)
r2$padj <- sample(r2$padj * .1)
VizGSEA(rbind(r, r2), group_by = "Group", plot_type = "heatmap")

VizGSEA(rbind(r, r2), group_by = "Group", plot_type = "heatmap", signif_only = FALSE)

VizGSEA(rbind(r, r2), group_by = "Group", plot_type = "dot")

VizGSEA(rbind(r, r2), group_by = "Group", plot_type = "dot", signif_only = FALSE)

# }