AI-Guided Precision in Antibody Humanization: Structural Modeling to Minimize Immunogenicity and Preserve Efficacy
Wang, H.; Xia, C.; Zhu, Y.; Wang, Y.; Chu, X.; Pan, X.; Zhou, H.; Lu, Z.; Patrick, D.; Li, Y.; Long, Y.; Yu, Q.
Show abstract
Antibody humanization is an essential part of converting animal-derived antibodies into clinical candidates however conventional CDR grafting techniques often faces challenges, especially when essential antigen-binding sites extend beyond CDR regions. Additionally, the CDR grafting process tends to retain undesired non-human elements within the CDR regions. In this study, we overcame these challenges by using our proprietary AI-predicted antigen-antibody complex algorithm to effectively humanize a murine antibody. During our AI-powered humanization process, in contrast to traditional CDR grafting techniques, only the crucial paratopes identified through our AI-predicted complex structure were precision grafted onto a select human germline. This paratope grafting technique significantly reduced the antibodies immunogenicity risk while maintaining bioactivity. Moreover, all AI-guided deep humanized antibody variants demonstrated favorable developability, including thermal stability, poly-reactivity, and hydrophobicity. These results not only advance antibody humanization techniques but also demonstrate the power of AI in expediting antibody engineering and derisking clinical research.
Matching journals
The top 10 journals account for 50% of the predicted probability mass.