Guaranteed Diversity and Optimality in Cost Function Network Based Computational Protein Design Methods

Proteins are the main active molecules of life. Although natural proteins play many roles, as enzymes or antibodies for example, there is a need to go beyond repertoire produce engineered that precisely meet application requirements, in terms function, stability, activity other protein capacities. Computational Protein Design aims at designing new from first principles, using full-atom molecular models. However, size and complexity require approximations make them amenable energetic optimization queries. These design process less reliable, provable optimal solution may fail. In practice, expensive libraries solutions therefore generated tested. this paper, we explore idea generating provably diverse low-energy by extending cost function network algorithms with dedicated automaton-based diversity constraints on large set realistic full redesign problems. We observe it possible generate reasonable time produced do enhance Native Sequence Recovery, traditional measure methods reliability.