How to add gene lable on GSEA plot

Question

[Emy Alade]

Hello, I performed a Gene Set Enrichment Analysis (GSEA) using the clusterProfiler package in R on a ranked gene list. I want to display the genes names directly on the enrichment plot (like in the example picture below).

library(readr)
library(ggplot2)
library(dplyr)
library(clusterProfiler)
library(biomaRt)
library(org.Hs.eg.db)
library(patchwork)
library(msigdbr)
library(enrichplot)

# Perform GSEA using gseGO function on the ranked gene list
gseGO_result <- gseGO(
                geneList = gene_list_unique,
                ont = "BP",                  # Use Biological Process ontology
                OrgDb = org.Hs.eg.db,        # Human gene annotation database
                keyType = "ENTREZID",        # Key type for gene IDs
                pvalueCutoff = 0.05,         # p-value cutoff for significant terms
                pAdjustMethod = "BH",        # Multiple testing correction method (Benjamini-Hochberg)
                minGSSize = 10,              # Minimum gene set size
                maxGSSize = 500,             # Maximum gene set size
                verbose = TRUE,
                seed = TRUE                  # Set seed for reproducibility
              )

# Simplify GSEA results by removing redundant GO terms based on similarity cutoff
gseGO_result2 <- clusterProfiler::simplify(
                  gseGO_result,
                  cutoff = 0.7,               # Similarity cutoff to merge redundant terms
                  by = "p.adjust",            # Use adjusted p-value for grouping
                  select_fun = min            # Select term with minimal adjusted p-value
                )

# Convert ENTREZ IDs to gene symbols for easier interpretation
gene_df <- bitr(
          names(gene_list_unique),
          fromType = "ENTREZID",
          toType = "SYMBOL",
          OrgDb = org.Hs.eg.db
        )

# Add gene symbols to the core enrichment column of the gseGO results
gseGO_result2@result$core_enrichment_symbol <- sapply(
  gseGO_result2@result$core_enrichment,
  function(x) {
    entrez_ids <- unlist(strsplit(x, "/"))
    symbols <- gene_df$SYMBOL[match(entrez_ids, gene_df$ENTREZID)]
    paste(symbols, collapse = "/")
  }
)


plot <- gseaplot2(gseGO_result2,
                  geneSetID = "GO:0071824",
                  pvalue_table = TRUE,
                  color =  "#86B875",
                 ES_geom = "line")

# Extraire les gènes "leading edge"
genes <- gseGO_result2@result[gseGO_result2@result$Description == "protein-DNA complex assembly", "core_enrichment"]
genes <- unlist(strsplit(genes, "/"))
deg_in_list <- sigDEGs_ids[sigDEGs_ids$ENTREZID %in% genes, ]
deg_in_list$rank <- match(deg_in_list$ENTREZID, names(gene_list_unique))
label_df <- na.omit(deg_in_list)
p3 <- plot[[3]]

# Préparer le dataframe pour les petites barres bleues à y=0
bars_df <- data.frame(
  x = label_df$rank,
  y = 0,
  xend = label_df$rank,
  yend = gene_list_unique[label_df$rank]
)

# Construire le plot
plot[[3]] <- p3 +  
  # Barres grises de tous les gènes
  geom_segment(data = data.frame(
                 x = seq_along(gene_list_unique),
                 y = 0,
                 xend = seq_along(gene_list_unique),
                 yend = gene_list_unique),
               aes(x = x, y = y, xend = xend, yend = yend),
               color = "grey70") +
  # Ligne horizontale à y=0
  geom_hline(yintercept = 0, linetype = "dashed") +
  # Petites barres bleues sur la baseline pour les gènes annotés
  geom_segment(data = bars_df,
               aes(x = x, y = y, xend = xend, yend = yend),
               color = "blue",
               size = 0.5) +
  # Texte annoté avec flèches partant de y=0 vers le texte au-dessus
  geom_text_repel(data = label_df,
                  aes(x = rank, y = 0, label = SYMBOL),
                  nudge_y = diff(range(gene_list_unique)) * 0.07,  # ajuster la hauteur du texte
                  angle = 90,
                  size = 3.5,
                  segment.size = 0.3,
                  segment.curvature = 0,
                  segment.ncp = 10,
                  segment.angle = 20) +
  labs(x = "Rank in ordered dataset", y = "Ranked list metric") +
  theme_minimal()

# Affichage avec patchwork, augmenter la hauteur du 3e plot si besoin
plot[[1]] / plot[[2]] / plot[[3]] + plot_layout(heights = c(3, 1, 3))

[yura.grabovska]

Hi Emy, it looks like from your worked example, you've managed to label some genes. What specifically would you like help with?

[Emy Alade]

I want the genes to be clearly readable and not overlapping (as in the first image). Normally, five genes should be displayed

[yura.grabovska]

You're more or less there...

Take a look at the ggrepel example below:
https://ggrepel.slowkow.com/articles/examples#limit-labels-to-a-specific-area

What I would recommend is that you select your genes of interest and feed them in via label_df then use a combination of geom_text_repel(xlim = ..., ylim = ..., nudge_x = ..., nudge_y = ...) to correctly define the regions that ggrepel is able to use. You may need to expand the y-axis limits for the plot go provide more space for the labels if you plan to use genes with very long names.

As an aside, I assume you want to code this rather than fix it in Illustrator because you plan to be generating many of these plots in a batch. I would still say there is an upper limit to how many of these plots are useful to show in this way. If you are going to be showing more than 3 or 4 of these, then maybe you could consider an alternative visualisation?

If you're able to share gene_list_unique, bars_df, label_df and gseGO_result2, or a dummy example of each I could help you with some code.