Automation and Active Learning for the Multi‐Objective Optimization of Antibody Formulations
Authors:
Open Access
- 13 July 2026
- journal article
- Published by Wiley in Advanced Science
- Vol. 13 (55) , e76551
Abstract
Over the last forty years, monoclonal antibodies have become increasingly important therapeutic agents, with most manufactured as preformulated solutions. However, bioformulation of complex proteins is a difficult engineering challenge; formulations must be tailored to individual therapies, necessitating time‐ and material‐intensive campaigns to select combinations of excipients to simultaneously optimize various design criteria. These additives complicate formulation design with unintuitive and non‐linear relationships, creating a vast multidimensional design space that is intrinsically difficult to optimize using traditional techniques. To address this challenge, we investigated a high‐throughput discovery pipeline using machine learning to model and predict the effects of GRAS excipients on formulation behavior of a model antibody. This was supported by automation‐assisted “on‐demand” formulation to produce dozens of uniquely formulated antibody solutions for downstream evaluation and biophysical characterization. This pipeline was then integrated into an iterative closed‐loop cycle of automated Design‐Build‐Test‐Learn (DBTL), where new rounds of experiments are designed by the model. The process yielded both improved formulations and accurate predictive models of formulation behavior across multiple target objectives (melting temperature, diffusivity, and high‐concentration viscosity). This validates the utility of this technique to both map the underlying property‐function landscape and effectively guide formulation development while balancing multiple competing design requirements.Keywords
All Related Versions
Funding Information
- National Institutes of Health (NIBIB R01EB037022)
- National Institutes of Health (NIGMS R35GM138296)
- National Science Foundation (DMREF 2118860)
- National Science Foundation (CBET 2309852)
This publication has 70 references indexed in Scilit:
- Bayesian cell therapy process optimizationBiotechnology and Bioengineering, 2024
- Machine Learning on a Robotic Platform for the Design of Polymer–Protein HybridsAdvanced Materials, 2022
- Machine‐Assisted Discovery of Chondroitinase ABC Complexes toward Sustained Neural RegenerationAdvanced Healthcare Materials, 2022
- Analysis of Protein Stability and Ligand Interactions by Thermal Shift AssayCurrent Protocols in Protein Science, 2015
- Comparative Effects of pH and Ionic Strength on Protein–Protein Interactions, Unfolding, and Aggregation for IgG1 AntibodiesJournal of Pharmaceutical Sciences, 2010
- High-throughput dynamic light scattering method for measuring viscosity of concentrated protein solutionsAnalytical Biochemistry, 2010
- High throughput thermostability screening of monoclonal antibody formulationsJournal of Pharmaceutical Sciences, 2010
- Kriging Is Well-Suited to Parallelize Optimization2010
- Antibody Structure, Instability, and FormulationJournal of Pharmaceutical Sciences, 2007
- On the distribution of points in a cube and the approximate evaluation of integralsUSSR Computational Mathematics and Mathematical Physics, 1967