Platform Science in Depth
Detailed scientific rationale, platform architecture, and academic partnership structures for each program.
ELEGANS-INTERACTOME
AI/ML Platform for Human Pathogen-Host Protein Interaction Discovery
The Science of Pathogen-Host Interactome
ELEGANS-INTERACTOME is a Graph Neural Network (GNN) platform that predicts direct, physical protein-protein interactions between human pathogens and their human host targets — from pathogen protein sequence alone, without requiring prior experimental structural data. The platform is built on a five-stage computational pipeline integrating sequence-to-structure prediction, spatial graph construction, dual-output GNN inference, and orthology-guided cross-species mapping from C. elegans to human cell validation.
C. elegans as an Experimental Discovery Engine
The platform is anchored in a 50-year scientific tradition of using Caenorhabditis elegans — the only animal whose complete nervous system, cell lineage, and genome have been fully mapped — as a model for human pathogen biology. C. elegans shares more than 60% of human disease-gene orthologues and has been validated as a model host for bacterial, fungal, and viral pathogens including Pseudomonas aeruginosa, Staphylococcus aureus, Candida albicans, and multiple emerging infectious agents. The speed, genetic tractability, and scalability of C. elegans infection models makes it uniquely suited to rapidly screen predicted pathogen-host interactions at scale before validation in human cells.
Collaborative Research Validation Network
McGenome is actively developing research collaborations with leading C. elegans pathogen biology laboratories and human cell validation groups to provide experimental grounding for ELEGANS-INTERACTOME predictions. These collaborations span institutions with established expertise in nematode host defense, innate immunity, and human pulmonary and vascular cell systems.
Target Platform Applications
Platform applications include:
- Identification of novel pathogen virulence factors with human protein targets, including those with no prior structural characterization
- Biodefense and biosurveillance: rapid characterization of emerging and engineered pathogen protein repertoires
- Drug target discovery: prioritization of pathogen-human interaction nodes for small molecule and biologic therapeutic development
- Antimicrobial resistance (AMR): pathway-level mapping of resistance-associated protein networks in priority pathogens
| Architecture | 3-layer Graph Attention Network (GAT) · 8 attention heads · 256 hidden dimensions · GNN/SpatialPPIv2 · Dual-output prediction head |
| Pipeline stages | Sequence input → Structure prediction → Graph construction → GNN inference → Orthology mapping → C. elegans / human cell validation |
| Inference latency | <60 seconds per pathogen protein query |
| Validation pipeline | C. elegans infection model → human cell confirmation → pathway multi-task annotation |
| Patent status | US PPA No. 64/096,508 · Filed June 23, 2026 · Micro Entity · Non-provisional deadline June 23, 2027 |
Platform details available to qualified partners under mutual NDA. Contact: ceo@mcgenome.com