Autonomous high-throughput biological screening for agrochemical discovery

Overview

This proposal describes an autonomous high-throughput biological screening platform designed to serve as the Test-Learn component of a closed-loop Design-Make-Test-Analyse (DMTA) cycle for agrochemical discovery. The system integrates automated liquid handling, low-volume enzyme activity and protein-ligand binding assays, plate-based measurements, and machine learning to screen compound libraries automatically, process results in real time, identify active compounds, generate concentration-response measurements, and prioritise subsequent compounds for testing.

The longer-term objective is to integrate this screening capability with automated molecular design, synthesis, and purification, enabling experimental biological data to directly inform the next molecules entering the discovery cycle. This would create a foundation for fully autonomous, end-to-end molecular discovery.

Technical specifications

Key features:

• Automated liquid handling for high-throughput compound screening
• Low-volume enzyme activity and protein-ligand binding assays
• Plate-based measurements with real-time data processing
• Machine learning to identify active compounds and prioritise next candidates
• Modular platform design supporting enzyme activity, inhibition, and binding assays
• Machine-readable biological results that directly inform subsequent molecular selection
• Integration with automated synthesis, purification, and analytical QC workflows

Planned development phases:

• Months 1-3: Adapt existing automated synthesis platform to a selected compound series, establishing automated reaction execution and sample/data tracking
• Months 3-6: Integrate automated purification and analytical QC, alongside direct-to-biology workflows where purification is unnecessary
• Months 5-8: Integrate automated liquid handling and plate-reader assays for biological testing
• Months 8-12: Demonstrate an end-to-end DMTA cycle with automated capture of chemical and biological data for the next design cycle

Technology readiness level

The platform is at an early development stage, building on existing automation capabilities in the proposing laboratory. The first year of work would focus on integrating biological screening into an existing automated synthesis platform and demonstrating an end-to-end DMTA cycle benchmarked against current workflows. A stretch goal includes using the resulting dataset with machine learning models to propose the next compounds for automated synthesis, closing the DMTA loop. The platform is designed for transfer to other targets and compound series after initial demonstration.