Sample-efficient Antibody Design through Protein Language Model for Risk-aware Batch Bayesian Optimization
Keywords: Antibody Design, Protein Language Mode, Batch Bayesian Optimization
Abstract: Antibody design is a time-consuming and expensive process that often requires1
extensive experimentation to identify the best candidates. To address this challenge,2
we propose an efficient and risk-aware antibody design framework that leverages3
protein language models (PLMs) and batch Bayesian optimization (BO). Our4
framework utilizes the generative power of protein language models to predict5
candidate sequences with higher naturalness and a Bayesian optimization algorithm6
to iteratively explore the sequence space and identify the most promising candidates.7
To further improve the efficiency of the search process, we introduce a risk-aware8
approach that balances exploration and exploitation by incorporating uncertainty9
estimates into the acquisition function of the Bayesian optimization algorithm.10
We demonstrate the effectiveness of our approach through experiments on several11
benchmark datasets, showing that our framework outperforms state-of-the-art12
methods in terms of both efficiency and quality of the designed sequences. Our13
framework has the potential to accelerate the discovery of new antibodies and14
reduce the cost and time required for antib
Submission Number: 5
Loading