gws_gena

Purpose

This task takes a DEG table, validates/normalizes an NCBI Entrez GeneID column, then:

1. converts Entrez IDs to KEGG gene IDs (hsa:1234, dme:xxxx, …) using KEGG REST
2. builds the species KEGG universe (genes ↔ pathways)
3. computes KEGG pathway enrichment (hypergeometric) + BH/FDR correction
4. renders Pathview PNGs for enriched pathways, supporting multiple fold-change columns on the same PNG (Pathview multi-state → .multi.png).

Requirements

• The analysis column must contain NCBI Entrez GeneIDs: digits only (1234, 318213, …) no symbols, no Ensembl IDs, no “GeneID:…” values like 1234.0 should be normalized to 1234 (the pipeline typically cleans this, but clean input is safer).
• digits only (1234, 318213, …)
• no symbols, no Ensembl IDs, no “GeneID:…”
• values like 1234.0 should be normalized to 1234 (the pipeline typically cleans this, but clean input is safer).
• If you start from other identifiers (Ensembl, symbols…), first run OmiX “Gene ID conversion” to obtain ENTREZGENE.

Input

A DEG table (CSV/TSV) containing:

• one Entrez GeneID column
• optional one or more fold-change columns

Parameters

• Organism (autocomplete): scientific name (e.g., Homo sapiens, Drosophila melanogaster, Saccharomyces cerevisiae). The task automatically maps it to the KEGG code (hsa, dme, sce, …).
• col_entrez: column name containing Entrez GeneIDs.
• foldchange_cols (optional): comma-separated list of fold-change columns to display in Pathview. Example: FoldChange,FoldChange 21 column → *.pathview.png≥2 columns → *.pathview.multi.png
• min_genes_required (internal/script): minimum number of valid Entrez genes required to proceed.

Outputs

1. pathways (ResourceSet): Pathview PNG images typical names: hsa00480.pathview.png (single FC) hsa00480.pathview.multi.png (multiple FC)
2. kegg_enrichment (Table): enrichment results with at least: pathway, pathway_name, overlap, pathway_size, query_size, universe_size, pvalue, padj

How overlap / pvalue / padj are computed

• Universe: all KEGG genes for the species (from link/pathway/<specie>).
• Query: your mapped genes intersecting the universe.
• For each pathway: K = pathway gene set size in universe n = query size k = overlap = number of query genes in the pathway
• pvalue: hypergeometric right-tail (probability of ≥k overlaps by chance).
• padj: Benjamini–Hochberg (FDR) correction across all tested pathways.

Why can a PNG be produced with no colored genes?

Even if a pathway is enriched, Pathview may show no coloring when:

• gene IDs don’t match what Pathview expects (wrong id type / wrong formatting)
• fold-change values become NA / non-numeric after parsing
• the map contains mostly compound/KO nodes rather than gene nodes relevant to your IDs

The enrichment table remains valid; coloring depends on ID ↔ node compatibility in Pathview.