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

Performers will develop an in-silico capability for mapping all protein-protein interactions (PPIs). The program is structured as a 30-month program with a 12-month Phase 1 and an 18-month Phase 2. Phase 1 focuses on generating an initial PPI capability capturing interactions in a lower dimensional space, targeting ~100,000 PPIs modeled over >10 conditions. Phase 2 focuses on scaling to a PPI hypercatalog of 10 million unique interactions modeled over >50 unique cellular contexts. Proposed solutions should address criteria such as decreasing dimensionality of PPI energy landscapes, scalable strategies for modeling PPIs across the human proteome, modeling multiple system sizes, capturing key binding variables, internal benchmarking, and modeling contextual perturbations. The work will be performed in VA.