🧬

Network Gene Essentiality Predictor

NGEP combines WikiPathways and Reactome biological network analysis with XGBoost machine learning to predict gene essentiality from network topology — cascade size, betweenness, sink reach and more.

2,079
WikiPathways GPML files
Multi-organism curated biological pathways across 40+ species
754
Reactome GPML files
Human-focused, expert-curated signalling & metabolic reactions
2,833
Total Pathways
Indexed Genes
19,278
Known Lethal Genes
Organisms Covered

Quick Gene Search

Search across all 2,833 pathways and get essentiality predictions instantly.

How It Works

1
Dual-Database Pathway Parsing

GPML files from WikiPathways (2,079) and Reactome (754) are parsed as directed gene interaction graphs.

2,079 WP pathways 754 Reactome pathways 20+ organisms
2
Network Feature Extraction

8 topological features computed per gene node from graph analysis.

Out-degree / In-degree
Cascade Size & Sink Reach
Betweenness / Closeness / Eigenvector
Relative Impact Score
3
XGBoost ML Prediction

A gradient-boosted classifier trained on DEG database lethal/non-lethal labels predicts gene essentiality scores.

Score ≥ 0.6 → Lethal
19,278 verified lethal genes
36,856 non-lethal genes
Pathway Coverage by Organism
Database Source Split