Towards deep learning models for target-specific antibody design

Abstract

Over the last few years, deep learning (DL) models have made a significant impact on the field of protein structure prediction.1 Recently DL models such as AlphaFold22 have achieved state-of-the-art performance on blind structure prediction of difficult protein targets. Additionally, several DL models have also been proposed for variant stability prediction (ESM-v13) and design of de novo proteins (trDesign4). Protein design is the inverse problem of protein structure prediction, i.e., the prediction of sequence given structure.