Abstract The demonstrated effectiveness and vast potential of large molecule therapeutics have driven biopharmaceutical companies to change the ways by which they discover and develop novel biologic therapies. While the conventional antibody discovery process via animal immunization and in vivo hypermutation can generate sufficient hits for functional screening, this approach can only cover a limited subset of sequence diversity. In silico, artificial intelligence (AI) driven methods, by contrast, have the potential to cover 2-3 orders of magnitude more sequence diversity compared to conventional methods and more rapidly produce a drug candidate with a superior safety and efficacy profile. However, this approach requires wet lab validation of predicted sequences as well as targeted data generation to improve AI-predicted rankings. Traditional wet lab methods for protein generation and characterization are expensive, time consuming, and prone to human error. These limitations restrict the overall potential benefit of the AI-driven antibody discovery process. To accelerate this process, which consists of a complex series of cycles spanning different functional teams, a substantial increase in the throughput of the end-to-end production and characterization workflow is established in-house such that 104 antibodies with property data can be generated and analyzed within a short period of time. BioMap has implemented a highly integrated and automated protein expression, purification, and characterization platform linked with a unified informatics system to eliminate as much manual operation as possible. This high-throughput robotic platform interweaves multiple procedures including plasmid preparation, cell dispensing, mammalian transfection, protein purification, and characterization, and it is coupled to our in-house informatics system and database. Empowered by state-of-the-art liquid handling system and well-established experimental protocols, the integrated facility is capable of the delivery of 103 protein samples from plasmids in ten days within a single batch. Benefiting from our robust transient expression platform, the production scale ranges from 0.5 to 30 mL with an average antibody yield of 300 mg/L. The high-throughput workflow brings a challenge to track the provenance of each protein and manage data flow across different stations. To address this bottleneck, a well-defined informatics platform and database have been developed to provide sample registration, interface with lab instruments, experimental process tracking, and automated data recording and analysis. The implementation of this system provides a massive amount of high-quality data available for AI training and validation with a very short turnover time and enables AI-driven development of next-generation biologics. Citation Format: Zhehao Xiong, Wei Jiang, Lijun Xia, Jianhua Huang, Cheng-chi Chao. Accelerating the drug discovery process with an automated high-throughput protein production and characterization platform for AI-driven antibody development of immunotherapy. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5323.
Read full abstract