Real-Time Pathogen-Host Interactome Prediction

The DPA26BZ03-DV014 project seeks to develop a real-time pathogen-host interactome prediction capability to rapidly characterize host–pathogen interactions from pathogen protein sequences. This will enable timely medical countermeasure prioritization and force health protection against novel biological threats. The system must accurately predict host-pathogen protein interactions across various pathogen classes, demonstrate zero-shot prediction on unseen pathogens, provide comprehensive functional annotation, generate ranked mechanistic hypotheses, and complete core predictions within 15 minutes. The project is soliciting Direct to Phase II proposals, requiring prior demonstration of feasibility including benchmark performance data, zero-shot validation, pipeline demonstration, and functional annotation capability. Phase II involves scaling and validating the computational pipeline, conducting experimental validation of novel predicted interactions, and delivering transition-ready software with drug repurposing capabilities. Dual-use applications include military biosurveillance and commercial drug discovery. The goal is to create an operationally deployable capability for rapid biological threat characterization.