Skip to contents

Visualize Top Network Enrichment Statistics

Source: R/visualize_topn.R
visualize_topn.Rd

This function generates a bar plot to visualize the top n network enrichment results for a specified category. The plot highlights the effect size (beta) and the false discovery rate (-log10(FDR)), enabling quick assessment of the most enriched terms or entities in the network.

Usage

visualize_topn(
  tkoi_list,
  category = "BiologicalProcess",
  top_n = 25,
  ranknorm = FALSE,
  lognorm = TRUE,
  high_color = "#FF5733",
  low_color = "#154360"
)

Arguments

tkoi_list

A tKOIList object containing network summary statistics.

category

A character string specifying the category to visualize. Default is "BiologicalProcess". Accepted values include:

  • "Anatomy"

  • "BiologicalProcess"

  • "CellType"

  • "CellularComponent"

  • "ClinicalLab"

  • "Complex"

  • "Compound"

  • "Disease"

  • "EC"

  • "Gene"

  • "MiRNA"

  • "MolecularFunction"

  • "Pathway"

  • "Protein"

  • "ProteinDomain"

  • "ProteinFamily"

  • "PwGroup"

  • "Reaction".

top_n

An integer specifying the number of top results to display. Default is 25.

ranknorm

A logical value indicating whether to apply inverse rank-based normalization to the beta values. Default is FALSE.

lognorm

A logical value indicating whether to apply log base-2 transformation to the beta values. Default is TRUE.

high_color

A character string specifying the high value color for the gradient scale representing -log10(FDR). Default is "#FF5733".

low_color

A character string specifying the low value color for the gradient scale representing -log10(FDR). Default is "#154360".

Value

A ggplot object representing the bar plot of top network enrichment statistics.

Details

The function performs the following steps:

  1. Takes the table for category from the network_summary_statistics slot of the tKOIList object and drops rows with a missing beta.

  2. Keeps the first top_n remaining rows (the tables are already ranked).

  3. Optionally applies transformations to the beta values:

    • If ranknorm is set to TRUE, applies inverse rank-based normalization.

    • If lognorm is set to TRUE, applies log base-2 transformation.

    • If both transformations are enabled, ranknorm is overridden and set to FALSE.

  4. Raises fdr and p_value values below the smallest positive FDR of the category to that value, so zero FDRs can be plotted on a log scale.

  5. Labels bars with the name column for "BiologicalProcess", "CellularComponent", "MolecularFunction", "Disease", and "Gene" (gene symbols), falling back to identifier for nodes without a name; other categories are labelled with identifier.

  6. Filters out entries with missing identifiers, keeps the first row per identifier (annotation joins can repeat a node), and fixes the plotting order.

  7. Creates a horizontal bar plot where:

    • The x-axis represents the (possibly transformed) effect size (beta).

    • The y-axis represents the identifiers (e.g., terms or entities).

    • The color gradient of the bars represents -log10(FDR), with customizable low and high colors.

See also

ggplot, arrange, filter

Examples

if (FALSE) { # \dontrun{
# Visualize the top 10 Biological Processes with default transformations and colors
plt = visualize_topn(tkoi_list, category = "BiologicalProcess", top_n = 10)
print(plt)

# Visualize the top 20 Genes with custom color gradient
plt = visualize_topn(tkoi_list, category = "Gene", top_n = 20, high_color = "#E74C3C",
low_color = "#3498DB")
print(plt)

# Visualize the top 15 Pathways without rank-based normalization or log transformation
plt = visualize_topn(tkoi_list, category = "Pathway", top_n = 15, ranknorm = FALSE, lognorm =
FALSE)
print(plt)

# Visualize the top 5 Diseases with rank-based normalization enabled
plt = visualize_topn(tkoi_list, category = "Disease", top_n = 5, ranknorm = TRUE, lognorm =
FALSE)
print(plt)
} # }