Search papers, labs, and topics across Lattice.
The paper introduces BoGA (Bayesian Optimization Genetic Algorithm), a novel framework for protein design that integrates a genetic algorithm for stochastic proposal generation within a Bayesian optimization loop. This approach leverages surrogate modeling to prioritize candidate sequences based on prior evaluations and predictions, enabling efficient exploration of the protein sequence space. Benchmarking on sequence and structure design tasks, along with the design of peptide binders against pneumolysin, demonstrates BoGA's ability to accelerate the discovery of high-confidence binders.
By fusing Bayesian optimization with a genetic algorithm, BoGA achieves data-efficient protein design, rapidly identifying high-affinity binders against pneumolysin.
Designing novel proteins with desired characteristics remains a significant challenge due to the large sequence space and the complexity of sequence-function relationships. Efficient exploration of this space to identify sequences that meet specific design criteria is crucial for advancing therapeutics and biotechnology. Here, we present BoGA (Bayesian Optimization Genetic Algorithm), a framework that combines evolutionary search with Bayesian optimization to efficiently navigate the sequence space. By integrating a genetic algorithm as a stochastic proposal generator within a surrogate modeling loop, BoGA prioritizes candidates based on prior evaluations and surrogate model predictions, enabling data-efficient optimization. We demonstrate the utility of BoGA through benchmarking on sequence and structure design tasks, followed by its application in designing peptide binders against pneumolysin, a key virulence factor of \textit{Streptococcus pneumoniae}. BoGA accelerates the discovery of high-confidence binders, demonstrating the potential for efficient protein design across diverse objectives. The algorithm is implemented within the BoPep suite and is available under an MIT license at \href{https://github.com/ErikHartman/bopep}{GitHub}.
