Catalysis Modeling
AI/ML in conjunction with QM for reaction modeling to decrease avoidable experiment expenditure
Context
Multiple rounds of HTE to optimize reaction conditions delays drug development
Identifying conditions that produce desired yields involves conducting multiple high-throughput experiments (HTEs).
Such iterative experimentation leads to escalating material and labor cost and delayed timelines for the Process R&D function.
On the other hand, decades of experimentation; especially on well-known reactions such as Suzuki coupling, and Buchwald-Hartwig amination; have built up a corpus of public and proprietary data.
Aganitha applies state-of-the-art AI/ML models to HTE data to build predictive and prescriptive models for yield prediction.
Our Solution
Customized model development to predict reaction yields using HTE data
With our Analytics & AI/ML platform, we can help you identify the right combination of base, ligand, and solvent in the very first iteration of the 96-well plate experiments. Additionally, we can help you with:
- Analytics tool to identify low-performing substrates and high-performing combinations.
- Domain-based physics descriptors to build explainable prediction models.
- UI tailored for chemists to rapidly ideate and identify experiments to be carried out in wet lab.
