UNMET NEED
• Protein-protein interactions (PPIs) are essential components of all cell signaling pathways and important therapeutic targets. However, the design of PPI modulators, especially in the form of small molecules, remains a major challenge.
• Interfering with PPIs is especially challenging when one of the interactors is a disordered protein (i.e., with no stable structure).
• Peptides, i.e., proteins of small length (L<30-40) are a promising class of PPI modulators. However, screening the vast sequence space (20L peptides of length L) is well-beyond the capabilities of experimental investigation and computational approaches based on molecular docking.

OUR SOLUTION
• We developed and published an integrative peptide design protocol for peptide binder discovery based on a machine learning sequence generative model1. 
• The generative model (a Restricted Boltzmann Machine) is trained on known and putative protein binders of the target protein identified from previous experiments and homology search on large sequence databases. The model “reverse-engineers” the key sequence motifs underlying binding, and generates novel sequences by recombination of these motifs.
• We tested our protocol on Calcineurin, an important target for immunosuppressive therapy, and discovered in a single screening round multiple novel peptides capable of interfering with its function. 
• Specifically, 70% of the designed peptides successfully interfered with Calcineurin-substrate interactions, including one with 10x higher binding affinity than the known PVIVIT peptide previously identified.
• Our solution is broadly applicable for designing inhibitors of PPIs where (1) One of the interactors is disordered and (2) Experimental structures, or accurate models of the complex exist.

APPLICATIONS
• Design of peptide inhibitors of protein-protein interactions.

STATUS
One article published; several ongoing follow-up projects.

INTELLECTUAL PROPERTY
Ongoing patent application.

REFERENCES
1. Tubiana, J., Adriana-Lifshits, L., Nissan, M., Gabay, M., Sher, I., Sova, M., … & Gal, M. (2023). Funneling modulatory peptide design with generative models: Discovery and characterization of disruptors of calcineurin protein-protein interactions. PLOS Computational Biology, 19(2), e1010874.