Using gene co-expression graph topological fingerprints to enhance pathway activity scoring via xGATE
Quantifying pathway activity in single-cell and spatial omics is error-prone when methods ignore how genes wire together. We present xGATE, a topology-aware pathway-scoring framework that reads activity from the topological fingerprints of gene co-expression graphs. xGATE summarizes graph structure into embeddings and calibrates significance against a random-subgraph null, yielding interpretable p -values and continuous effect sizes. Across diverse omics datasets, xGATE outperforms marginal-expression and correlation-counting methods, recovering known active pathways and capturing dynamics that elude gene markers. xGATE yields mechanistic pathway readouts across tissues and platforms.
Authors
- Orlando Ferrer
- Yuxia Xie
- Jichun Xie
Institutions
- Duke University (US)
Publication Details
- Journal
- Genome biology
- Published
- 2026-11-21
- DOI
- https://doi.org/10.1186/s13059-026-04341-2
- Primary Topic
- Bioinformatics and Genomic Networks
- Type
- article
- Field-Weighted Citation Impact
- 0.00