OMIX GSEA

Gene Set Enrichment Analysis (GSEA) for pre-ranked gene lists using the fgsea package.

Canonical OMIX module

Item	Location
Canonical module	OMIX GSEA Preranked Legacy
Interface contract	schemas/interface.yml
Development contract	OMIX module contract
Released source reference	OMIX_MODULE_SOURCE.md

The canonical module owns scientific behavior, portable CLI operation, tests, and the reusable input/output contract. This repository is its Code Ocean deployment adapter.

What this deployment adds

• The Code Ocean App Panel and capsule entry point.
• Uploaded and attached DEG-table discovery under /data.
• Capsule outputs under /results and the pinned runtime configuration.

Overview

Version: v5.0 | Part of the OMIX analysis suite

Gene Set Enrichment Analysis (GSEA) determines whether predefined gene sets show statistically significant, concordant differences between two biological states. This tool uses pre-ranked gene lists from differential expression analysis to evaluate pathway enrichment using the fast fgsea algorithm.

What is GSEA?

Unlike over-representation analysis (ORA) which tests if specific genes are enriched in a pathway, GSEA evaluates whether genes in a pathway are concentrated at the top or bottom of a ranked gene list. This approach:

• Uses all genes in your data, not just significant ones
• Considers gene rankings (t-statistics, fold changes), preserving biological signal
• Identifies subtle but coordinated changes in pathway activity
• Is more sensitive to pathway-level effects than ORA

Key Features

• Pre-ranked GSEA: Uses your gene rankings (t-statistics, log fold-change) directly
• Comprehensive MSigDB coverage: Access to Hallmark, GO, REACTOME, KEGG, and more
• Cross-species support: Automatic ortholog mapping for non-human organisms
• Multiple contrasts: Analyze multiple comparisons in a single run
• Fast permutation testing: Efficient multi-level split Monte-Carlo scheme
• Leading edge genes: Identifies core genes driving enrichment signals
• Publication-ready outputs: Enrichment statistics, FDR correction, and plots

How It Works

1. Input: Ranked gene list from differential expression (e.g., sorted by t-statistic)
2. Enrichment Calculation: For each pathway, calculates an enrichment score (ES) that reflects whether pathway genes are concentrated at the top (upregulated) or bottom (downregulated) of the ranking
3. Statistical Testing: Permutation testing determines if ES is significantly different from random
4. FDR Correction: Benjamini-Hochberg correction controls for multiple testing
5. Output: Normalized enrichment scores (NES), p-values, FDR, and leading edge genes

When to Use GSEA vs. ORA

Use GSEA when:

• You want to detect subtle, coordinated pathway changes
• You have well-powered experiments with reliable gene rankings
• You want to use information from all genes, not just significant ones
• You're analyzing bulk RNA-seq with good statistical power

Use ORA (L2P tools) when:

• You have a clear set of differentially expressed genes
• You want simpler, more interpretable results
• You're working with heterogeneous data (e.g., single-cell)
• You want to compare enrichment across multiple conditions side-by-side

Input Requirements

DEG Table (Data Asset or Upload)

Differential expression results with:

• Gene identifiers (e.g., HGNC symbols)
• Ranking scores: t-statistics or log fold-changes for each comparison
• Format: RDS, CSV, or TSV

Column naming pattern: {comparison}_{suffix}
Example: TreatmentA_tstat, TreatmentB_tstat, Control_tstat

Contrast selection and order

Leave Contrasts (Optional) blank to analyze every ranking-score column that ends in the selected Gene Score Suffix, in the input-table column order. To run a selected ordered subset, enter comma-separated names such as B-A,C-A,C-B. Each name must exactly match the prefix of a ranking-score column (for example, B-A_tstat).

Demo dataset included: Mouse RNA-seq data runs automatically if no file is uploaded.

Supported Species

Human, Mouse, Rat, Dog, Rabbit, Zebrafish, Drosophila, Chimpanzee, Macaque

Non-human genes are automatically converted to human orthologs before pathway analysis, then results are mapped back to original gene IDs.

Pathway Collections

Access to MSigDB collections including:

• Hallmark (H): Core biological states and processes (50 gene sets)
• Curated pathways (C2): REACTOME, KEGG, BioCarta, WikiPathways
• Regulatory targets (C3): Transcription factors, microRNAs
• Gene Ontology (C5): Biological process, cellular component, molecular function
• Oncogenic (C6): Cancer gene signatures
• Immunologic (C7): Immune cell states and perturbations
• Cell type (C8): Cell type marker genes

Recommendation: Start with Hallmark for the most interpretable results.

Output Files

1. gsea_results.csv

Complete enrichment results table:

• Pathway name and collection
• NES (Normalized Enrichment Score): Positive = enriched in upregulated genes, negative = downregulated
• P-value and FDR: Statistical significance
• Gene set size and overlap
• Leading edge genes: Core genes driving enrichment

2. Plots (if generated)

• Enrichment plots showing pathway distribution across gene rankings
• Summary tables of significant pathways

Key Result Metrics

• NES (Normalized Enrichment Score): Effect size; |NES| > 1.5 typically indicates strong enrichment
• P-value: Raw statistical significance from permutation test
• FDR (False Discovery Rate): Multiple testing-corrected p-value; < 0.05 is typically significant
• Leading Edge Genes: Subset of pathway genes contributing most to enrichment

Algorithm Details

Uses the fgsea R Bioconductor package:

• Fast pre-ranked GSEA: Adaptive multi-level split Monte-Carlo permutation scheme
• Permutation testing: Estimates significance while preserving gene-gene correlations
• FDR correction: Benjamini-Hochberg procedure controls false discovery rate
• Normalized scoring: Accounts for pathway size differences

Method Comparison

Feature	GSEA (this tool)	ORA (L2P tools)
Input	Ranked gene list	Gene subset (DEGs)
Uses all genes	✓	✗
Statistical power	Higher for subtle effects	Higher for strong effects
Interpretation	Enrichment score + direction	Hit count + odds ratio
Multi-contrast viz	Table	Bubble plot

References

• fgsea: Korotkevich G et al. (2021). Fast gene set enrichment analysis. bioRxiv. doi:10.1101/060012
• Original GSEA: Subramanian A et al. (2005). Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. PNAS. doi:10.1073/pnas.0506580102
• MSigDB: Liberzon A et al. (2015). The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Systems. doi:10.1016/j.celsys.2015.12.004

Support

For questions or issues:

• NIDAP Team: NCICCBRNIDAP@mail.nih.gov
• CCBR: CCR Collaborative Bioinformatics Resource

For detailed step-by-step instructions on using the App Panel, see /code/README.md.

For developers

Read AGENTS.md and OMIX_MODULE_SOURCE.md before editing. Reusable scientific changes belong in the canonical OMIX module and are exported here only after validation.

---

OMIX Collection | Bioinformatics | Pathway Analysis | R | GSEA