Network-Based Gene Set Enrichment Analysis
Also known as: network GSEA, network-propagation GSEA, NetGSA, graph-informed gene set testing
Network-based gene set enrichment analysis (network GSEA) extends classical GSEA by incorporating biological interaction networks — such as protein-protein interaction (PPI) or co-expression graphs — into the enrichment test. Instead of treating each gene independently, the method propagates differential expression signals across network edges, allowing genes that are co-regulated or functionally connected to jointly support the significance of a gene set. The result is a biologically coherent enrichment score that accounts for pathway topology and gene-gene dependencies.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use network-based GSEA when gene expression changes are expected to be modest and distributed across a functionally coherent subnetwork rather than concentrated in a few highly significant genes. It is especially valuable in complex disease studies where pathway crosstalk is prominent, and in experiments where classical GSEA returns few significant gene sets despite a biologically meaningful treatment. A high-quality, species-appropriate interaction network must be available; do not use this method when no suitable network exists or when the network is sparse and poorly characterised for the organism of interest. It is also not appropriate as a first-line method in large-effect experimental comparisons where classical GSEA already produces clear results — the added network complexity may then obscure interpretation.
Strengths & limitations
- Captures pathway-level signal even when individual gene fold-changes are small, increasing sensitivity in complex disease contexts.
- Biologically principled: exploits known gene-gene interactions rather than treating the transcriptome as an unstructured list.
- Reduces noise from isolated outlier genes by requiring network-level corroboration of differential signals.
- Compatible with diverse network types (PPI, co-expression, regulatory) enabling flexible study design.
- Can identify novel pathway crosstalk by revealing enriched subnetworks that span multiple classical pathways.
- Results are directly dependent on the quality, completeness, and species-relevance of the interaction network used.
- Network propagation introduces an additional layer of modelling assumptions that can be difficult to validate or communicate to non-specialist audiences.
- Computationally more demanding than classical GSEA, especially when permutation-based null distributions must respect network topology.
- Less interpretable to non-specialists because the propagation step makes it non-obvious why a particular gene receives an elevated score.
Frequently asked
How is network GSEA different from classical GSEA?
Classical GSEA ranks genes by their individual expression statistics and tests whether a gene set is non-randomly distributed in that ranked list. Network GSEA first propagates expression signals across a biological interaction network so that genes connected to dysregulated neighbours receive higher scores even if their own fold-change is modest. The enrichment test then operates on these network-smoothed scores, making it more sensitive to coordinated but individually weak pathway perturbations.
Which network should I use?
Choose a network that matches your organism, tissue type, and biological question. STRING is widely used for broad PPI coverage; BioGRID for experimentally validated interactions; KEGG or Reactome for canonical signalling; co-expression networks (e.g., WGCNA-derived) for tissue-specific functional relationships. Avoid mixing networks from different organisms or using a pan-tissue network for a highly tissue-specific experiment without filtering.
What tools implement network-based GSEA?
NetGSA (R package on Bioconductor) implements the Shojaie-Michailidis linear model approach. EnrichNet provides network-distance-based enrichment via a web interface. HotNet2 identifies significantly dysregulated subnetworks. NEAT and the graphite R package offer network-topology-aware pathway testing. The choice depends on whether you need a formal statistical test framework (NetGSA), subnetwork discovery (HotNet2), or a topology-sensitive overlap test (NEAT).
Does network GSEA require a larger sample size than classical GSEA?
Not inherently, but the power gains from network propagation are most pronounced in larger datasets where the smoothing step has enough signal to propagate meaningfully. In very small experiments (fewer than five or six samples per group), both classical and network GSEA will have low power; network propagation does not substitute for adequate sample size.
Can I apply network GSEA to single-cell RNA-seq data?
Yes, but with additional considerations. Single-cell data typically requires pseudo-bulk aggregation or cell-type-specific differential expression before applying network GSEA in its classical form. Alternatively, cell-type-specific co-expression networks constructed from the single-cell data can serve as the propagation graph, enabling cell-type-resolved pathway enrichment.
Sources
- Shojaie, A., & Michailidis, G. (2010). Network enrichment analysis in complex experiments. Statistical Applications in Genetics and Molecular Biology, 9(1), Article 22. link ↗
- Subramanian, A., Tamayo, P., Mootha, V. K., Mukherjee, S., Ebert, B. L., Gillette, M. A., Paulovich, A., Pomeroy, S. L., Golub, T. R., Lander, E. S., & Mesirov, J. P. (2005). Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proceedings of the National Academy of Sciences, 102(43), 15545-15550. link ↗
How to cite this page
ScholarGate. (2026, June 3). Network-Based Gene Set Enrichment Analysis. ScholarGate. https://scholargate.app/en/bioinformatics/network-based-gene-set-enrichment-analysis
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Gene Set Enrichment AnalysisBioinformatics↔ compare
- Network-based GWASBioinformatics↔ compare
- Pathway Enrichment AnalysisBioinformatics↔ compare
- RNA-seq Differential ExpressionBioinformatics↔ compare
- Single-cell Gene Set Enrichment AnalysisBioinformatics↔ compare