(543g) Accelerating Drug Development with Suntheticsml: Removing Barriers for Machine Learning Adoption in Reaction Optimization

The development of new molecules, formulations, and scale-up processes in the pharmaceutical industry requires complex optimizations, which are often time-consuming, expensive, and too reliant on intuition and past experience. Machine-learning (ML) technologies have the potential to improve efficiency and streamline process development. However, only 33% of chemical companies currently use AI for the design of experiments and process optimization due to challenges in accessing and implementing ML tools.¹ These challenges include the lack of necessary coding and software skills among R&D scientists² and the lack of efficient algorithms designed to address complex chemistry optimization problems. SuntheticsML removed these barriers by incorporating our proprietary machine-learning algorithms into a user-friendly, code-free, online software platform specifically designed for R&D scientists to guide them during optimization campaigns.

It typically takes about 12-15 years to take a new medication from candidate selection to launch.³ As a result, pharmaceutical companies try to optimize their development process to minimize both the cost per asset and time to market. However, their efforts have primarily focused on clinical trials rather than preclinical development, which involves the period from candidate selection to the first human studies (FIH). Consequently, optimizing preclinical development presents a considerable untapped opportunity for these companies. Preclinical studies take an average of 21-26 months, of which 8-11 months are spent on process development to generate enough material for the following test.⁴ This stage of process development has an average cost of $ 1.5 M per drug candidate, just in personnel costs.⁵ According to a 2023 McKinsey report,⁴ pharmaceutical companies can reduce the time required to reach FIH applications by 40% or more by using the right process optimization technique. This not only provides patients with more rapid access to innovative medicines, but also gives companies earlier revenue flows and a longer period of exclusivity in the market. If a pharmaceutical company plans to advance three to five investigational new drugs into FIH studies each year, accelerating the Preclinical development by nine to 12 months for the entire portfolio could yield a risk-adjusted net present value of more than $400 million.⁴

Using an enhanced Bayesian Optimization approach, we elucidated a reaction-agnostic methodology for the design and implementation of intelligent experimental campaigns that leverages small data (starting with as little as 5 data points) and reaches unprecedented efficiencies saving up to 15x less time and experiments. SuntheticsML can be used in multi-target optimization of numerical and categorical variables and can be seamlessly integrated with any software or hardware platform for reaction optimization. In this talk, we will showcase recent case studies featuring a range of pharmaceutically relevant reactions, including catalysis and electrochemistry. We will highlight the power of Bayesian optimization and chemical knowledge to accelerate process development using small data to fast-track innovation, sustainability, and digitalization in pharmaceuticals, further reducing development waste, emissions, and resource consumption by up to 95%.

References:

(1) Krishnan, V.; Womack, D.; Lin, S. Optimizing the Chemicals Value Chain with AI | IBM; IBM, 2020. https://www.ibm.com/thought-leadership/institute-business-value/report/c... (accessed 2023-02-24).

(2) OECD. Artificial Intelligence and Machine Learning in Science; OECD: Paris, 2018; pp 121–136.

(3) Hughes, J.; Rees, S.; Kalindjian, S.; Philpott, K. Principles of Early Drug Discovery. Br. J. Pharmacol. 2011, 162 (6), 1239–1249.

(4) Agrawal, G.; Bader, F.; Günthner, J.; Wurzer, S. Fast to First-in-Human: How Preclinical Development Can Get New Medicines to Patients More Quickly | McKinsey; McKinsey & Company, 2023. https://www.mckinsey.com/industries/life-sciences/our-insights/fast-to-f... (accessed 2023-02-20).

(5) Farid, S. S.; Baron, M.; Stamatis, C.; Nie, W.; Coffman, J. Benchmarking Biopharmaceutical Process Development and Manufacturing Cost Contributions to R&D. mAbs 12 (1), 1754999.