Yakov Kipnis

RFdiffusion2 scaffolds unindexed atom-level catalytic active sites via flow matching without inverse rotamer enumeration, solving all 41 in silico benchmark cases and yielding active de novo enzymes across four reactions.

An atom-level graph neural network that denoises corrupted structures generates small-molecule and protein-ligand conformational ensembles, and its active-site pre-organization metric enriched for more active designed retroaldolases.

This study designs novel serine hydrolases with high catalytic efficiency using computational methods, achieving structures close to design models and distinct from natural enzymes.

This paper introduces ProteinMPNN, a deep learning model that improves protein expression, stability, and function by incorporating evolutionary and structural information.