# Enhancing antibody-antigen interaction prediction with atomic flexibility

## TL;DR

This paper shows how using predicted flexibility scores improves antibody-antigen interaction modeling, leading to better predictions and design of antibodies for vaccines and therapies.

## Contribution

The study introduces a novel fingerprint-based approach using pLDDT scores to model antibody flexibility, improving Ab-Ag interaction prediction accuracy by 4%.

## Key findings

- Incorporating pLDDT scores as a proxy for flexibility improved Ab-Ag interaction prediction with an AUC-ROC of 92%.
- The method achieved state-of-the-art performance in paratope prediction.
- Greater antibody flexibility enhances tolerance to antigen sequence variations, aiding in targeting pathogens like HIV and SARS-CoV-2.

## Abstract

Antibodies are indispensable components of the immune system, known for their specific binding to antigens. Beyond their natural immunological functions, they are fundamental in developing vaccines and therapeutic interventions for infectious diseases. The complex architecture of antibodies, particularly their variable regions responsible for antigen recognition, presents significant challenges for computational modeling. Recent advancements in deep learning have markedly improved protein structure prediction; however, accurately modeling antibody-antigen (Ab-Ag) interactions remains challenging due to the inherent flexibility of antibodies and the dynamic nature of binding processes. In this study, we examine the use of predicted Local Distance Difference Test (pLDDT) scores as indicators of residue and side-chain flexibility to model Ab-Ag interactions through a fingerprint-based approach. We demonstrate the significance of flexibility in different antibody-specific tasks, enhancing the predictive accuracy of Ab-Ag interaction models by 4%, resulting in an AUC-ROC of 92%. In addition, we showcase state-of-the-art performance in paratope prediction. These results emphasize the importance of accounting for conformational flexibility in modeling antibody-antigen interactions and show that pLDDT can serve as a coarse proxy for these dynamic features. By optimizing antibody flexibility using pLDDT, they can be engineered to improve affinity or breadth for a specific target. This approach is particularly beneficial for addressing highly variable pathogens like HIV and SARS-CoV-2, as greater flexibility enhances tolerance to sequence variations in target antigens.

Antibodies are crucial immune molecules that bind to antigens with high specificity, playing a central role in host defense, vaccine development, and therapeutic interventions. However, accurately modeling antibody-antigen (Ab-Ag) interactions presents significant challenges due to the intrinsic structural flexibility of antibodies, particularly in the complementarity-determining region H3 (CDRH3) and associated side chains. This study examines the impact of flexibility on antibody-antigen interactions by employing a fingerprint-based approach that incorporates ESMFold confidence scores as a proxy for antibody flexibility. We achieved a 4% improvement in predictive accuracy for Ab-Ag interactions, resulting in an AUC-ROC of 92% and demonstrated state-of-the-art performance in the prediction of paratopes. These findings underscore the importance of incorporating pLDDT in the prediction and design of antibody-antigen interactions.