Complementarity of Deep Learning and Physics-Based Approaches in the Design of New Antibodies
摘要
Antibodies and antibody derivatives have become the main focus in recent drug discovery due to their high specificity and affinity toward their targets. These biological agents offer a diverse range of characteristics and properties, rendering them interesting for drug development and engineering. In the last decade, approximately 175 antibody-based therapeutics have been under regulatory review or approved. These antibodies were engineered using various approaches, spanning from directed evolution to in silico drug design. In this chapter, we will mainly focus on utilizing in silico tools, such as physics-based approaches and artificial intelligence tools to understand the folding of antibodies and antibody fragments such as nanobodies, to decipher the antibody–antigen interactions, and to predict antibody properties. These aspects are essential for the engineering of antibodies and their derivatives.