Deep learning guided high-throughput virtual screening for in vitro antibody maturation

Antibody affinity maturation is a crucial step in therapeutic antibody discovery. In this study, we present a virtual screening pipeline that integrates protein docking with deep learning-based structural prediction to identify antibody mutants with enhanced binding affinity for the antigen sST2. By...

Full description

Saved in:
Bibliographic Details
Main Authors: Gong Chen, Liu Hongde
Format: Article
Language:English
Published: EDP Sciences 2025-01-01
Series:BIO Web of Conferences
Online Access:https://www.bio-conferences.org/articles/bioconf/pdf/2025/25/bioconf_icbb2025_03017.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850152343459332096
author Gong Chen
Liu Hongde
author_facet Gong Chen
Liu Hongde
author_sort Gong Chen
collection DOAJ
description Antibody affinity maturation is a crucial step in therapeutic antibody discovery. In this study, we present a virtual screening pipeline that integrates protein docking with deep learning-based structural prediction to identify antibody mutants with enhanced binding affinity for the antigen sST2. By introducing random mutations in the CDR3 domain of the wild-type antibody 2B4, we generated 949 mutants and systematically narrowed them down to 14 candidates for experimental validation. Among these, six exhibited higher affinity, while three displayed comparable affinity to 2B4. Notably, the selected mutants shared close interaction sites with each other, providing valuable region for antibody engineering and therapeutic development. Our pipeline is easy for local deployment with less computational resources required, offering a convenient tool for in-silico screening without the need of intensive cluster. The use of advanced deep learning structure prediction model further enhances the accuracy compared to traditional virtual screening pipeline. Consequently, our work significantly reduces the cost and time required for experimental in vitro affinity maturation, effectively combining the interpretability of protein docking with the predictive accuracy of deep learning.
format Article
id doaj-art-e6b6be88524a49659f9d25de79596331
institution OA Journals
issn 2117-4458
language English
publishDate 2025-01-01
publisher EDP Sciences
record_format Article
series BIO Web of Conferences
spelling doaj-art-e6b6be88524a49659f9d25de795963312025-08-20T02:26:01ZengEDP SciencesBIO Web of Conferences2117-44582025-01-011740301710.1051/bioconf/202517403017bioconf_icbb2025_03017Deep learning guided high-throughput virtual screening for in vitro antibody maturationGong Chen0Liu Hongde1State Key Laboratory of Digital Medical Engineering, School of Biological Science & Medical Engineering, Southeast UniversityState Key Laboratory of Digital Medical Engineering, School of Biological Science & Medical Engineering, Southeast UniversityAntibody affinity maturation is a crucial step in therapeutic antibody discovery. In this study, we present a virtual screening pipeline that integrates protein docking with deep learning-based structural prediction to identify antibody mutants with enhanced binding affinity for the antigen sST2. By introducing random mutations in the CDR3 domain of the wild-type antibody 2B4, we generated 949 mutants and systematically narrowed them down to 14 candidates for experimental validation. Among these, six exhibited higher affinity, while three displayed comparable affinity to 2B4. Notably, the selected mutants shared close interaction sites with each other, providing valuable region for antibody engineering and therapeutic development. Our pipeline is easy for local deployment with less computational resources required, offering a convenient tool for in-silico screening without the need of intensive cluster. The use of advanced deep learning structure prediction model further enhances the accuracy compared to traditional virtual screening pipeline. Consequently, our work significantly reduces the cost and time required for experimental in vitro affinity maturation, effectively combining the interpretability of protein docking with the predictive accuracy of deep learning.https://www.bio-conferences.org/articles/bioconf/pdf/2025/25/bioconf_icbb2025_03017.pdf
spellingShingle Gong Chen
Liu Hongde
Deep learning guided high-throughput virtual screening for in vitro antibody maturation
BIO Web of Conferences
title Deep learning guided high-throughput virtual screening for in vitro antibody maturation
title_full Deep learning guided high-throughput virtual screening for in vitro antibody maturation
title_fullStr Deep learning guided high-throughput virtual screening for in vitro antibody maturation
title_full_unstemmed Deep learning guided high-throughput virtual screening for in vitro antibody maturation
title_short Deep learning guided high-throughput virtual screening for in vitro antibody maturation
title_sort deep learning guided high throughput virtual screening for in vitro antibody maturation
url https://www.bio-conferences.org/articles/bioconf/pdf/2025/25/bioconf_icbb2025_03017.pdf
work_keys_str_mv AT gongchen deeplearningguidedhighthroughputvirtualscreeningforinvitroantibodymaturation
AT liuhongde deeplearningguidedhighthroughputvirtualscreeningforinvitroantibodymaturation