High-throughput Unmasking of Biological Interactomes with Binding Landscape Emulators (HUBBLE)

The DARPA HUBBLE program seeks innovative, high‑throughput computational and experimental approaches to map the full human protein‑protein interaction (PPI) landscape and apply it to traumatic‑brain‑injury (TBI) research. Phase I (12 months) requires creation of an initial interactome (~100 000 PPIs across >10 biological contexts), validation of a subset of predictions, and delivery of a reduced‑dimensional “PPI signature” that captures binding energy, residence time, and contact residues. Proposers must demonstrate capability in large‑scale AI‑driven modeling, experimental validation, and a clear benchmarking strategy, while complying with DARPA’s CMMC Level 1, data‑rights, and foreign‑national requirements. • Generate ~100 k PPI predictions for TBI‑related proteins across multiple cellular contexts. • Validate a representative subset in vitro/in vivo using DARPA‑provided test sets. • Deliver a signature‑based, low‑dimensional representation of each interaction (binding energy, residence time, contact residues). • Meet quantitative technical criteria (e.g., >60 % accuracy, ipTM >0.7, >10 contexts) and provide a detailed milestone‑payment schedule.