Automated stopped-flow library synthesis for rapid optimisation and machine learning directed experimentation.

Area of Science:

• Chemical Synthesis
• Drug Discovery
• Automation and Machine Learning

Background:

• Library synthesis is crucial for identifying new drug candidates, with library size, diversity, and synthesis time being key factors.
• Traditional batch and continuous flow chemistry methods have limitations in optimizing these parameters for pharmaceutical discovery.
• There is a need for intermediate synthesis approaches that balance automation, reagent efficiency, and reaction control.

Purpose of the Study:

• To introduce and evaluate stopped-flow synthesis as an efficient alternative for small molecule pharmaceutical discovery.
• To integrate a stopped-flow reactor into a high-throughput platform for automated combinatorial library synthesis and analysis.
• To demonstrate the synergy of experimental automation with machine learning for optimizing drug discovery processes.

Main Methods:

• Integration of a stopped-flow reactor into a high-throughput continuous platform for automated experimentation.
• Synthesis of combinatorial libraries using stopped-flow chemistry with at-line reaction analysis.
• Development and application of a feed-forward neural network (FFNN) model trained on experimental data for predictive guidance.

Main Results:

• Approximately 900 reactions were completed in 192 hours, significantly accelerating the synthesis timeframe.
• The stopped-flow approach utilized approximately 10% of the reactants and solvents compared to a fully continuous method.
• The FFNN model demonstrated good predictability and success when tested against an external dataset, validating its predictive capabilities.

Conclusions:

• Stopped-flow synthesis provides an efficient, automated, and reagent-sparing method for combinatorial library synthesis in drug discovery.
• Combining automated experimentation with machine learning enhances experimental optimization and predictive accuracy within the drug discovery cycle.
• This integrated approach streamlines the design, make, test, and analysis (DMTA) cycle, leading to more efficient drug discovery investigations.