AntibodyGPS®- Multiparametric Optimization Of Therapeutic Antibodies Using Precise Characterization And Machine Learning

By Jennifer A. Codding-Bui, Kaare Bjerregaard-Andersen*, Cyrill Brunner**, Thomas J. Purcell, Divya T. Vavilala, Claes Gustafsson, and Mark Welch

ATUM has developed a platform that uses ML-guided GPS Engineering to optimize the development of antibody-based drugs. The AntibodyGPS^® platform combines gene design, de novo gene synthesis, and cloning with high throughput transient antibody expression and purification, as well as high-resolution antibody analytics. It also uses machine learning-based gene design to evaluate the impact of sequence changes. This study showcases how using this platform impacts all stages of an antibody's life cycle, from discovery to manufacturing. Explore the information and results collected when AntibodyGPS^® was applied to engineer a lead molecule targeting a mechanism in neuronal disease. The process involved selecting key quality attributes of the molecule as parameters in the engineering process, and using a design of experiments approach to construct a small set of antibody variants. The platform was able to create a multifunctional molecule meeting all functional and developability needs.