(39a) Unveiling the Potential of ML and Small Data in Pharmaceutical Process Development with Suntheticsml

Pharmaceutical process and product development present significant challenges in terms of cost, efficiency, and sustainability. Harnessing the power of data-driven machine learning (ML) to predict behavioral trends, variable effects, and optimal conditions offers a promising solution to fast-track innovation. However, conventional ML algorithms often require extensive datasets, posing a resource-intensive challenge in experimental data collection.

To address this bottleneck, we introduce an enhanced Bayesian optimization (BO) approach, enabling a reaction-agnostic pathway for designing and implementing intelligent experimental campaigns. SuntheticsML is an accessible online ML platform tailored for researchers without coding or ML expertise. The pivotal in-lab validation of this approach demonstrates compelling returns on material and experimental efficiency, as well as performance gains against a competitive baseline.

SuntheticsML is a versatile technology that allows numeric, discrete, and mixed-integer optimization problems with up to 20 input parameters. This BO-powered approach allows flexible execution in serial or parallel experimentation. Furthermore, it facilitates bounded-target, multi-objective, and constrained-input optimizations, enabling simultaneous enhancements in cost and material efficiency. The versatility of SuntheticsML is exemplified through case studies covering chemocatalytic reactions, biocatalytic cascades, and in-vitro mRNA transcription processes, among others.

The in-lab validation of SuntheticsML convincingly demonstrates impressive returns on material efficiency, with up to a 75% reduction in the use of expensive or complex reagents. Experimental efficiency sees notable gains, with a 2-6X reduction in optimization experiments. Moreover, the platform enables a 9-12% increase in previously-optimized yields.