Ponnapati, M., Sinha, S., Lynch, B., Boyden, E. and JACOBSON, J., CAMP: COMBINATORIAL ENGINEERING OF PROTEINS. In ICLR 2025 Workshop on Generative and Experimental Perspectives for Biomolecular Design.

Abstract

Protein recombination has long been a key method in protein engineering to diversify and optimize sequences. We enhance and evolve this approach by using a protein language model, where we found that when log likelihood in the language model is represented as a spline, abrupt transitions in the spline identify crossover sites for designing recombinant protein libraries. We use these sites to guide recombination of sequence blocks from evolutionarily related sequences using MCMC sampling. Language models also enable generation of novel recombinant blocks beyond traditional MSAs increasing diversity, while a direct preference optimization algorithm is used to fine-tune these blocks for reduced immunogenicity. This method integrates modern deep learning architectures with traditional protein engineering techniques to improve success rate of the libraries designed for wetlab verification.