Automated drug discovery using AI and robotics

AI + Medicine Newsletter 2024–01–16

Last year, researchers from Google DeepMind and Carnegie Mellon showcased the automatic materials discovery and chemical research using AI and robotics. This year, the AI-guided automated experiment is also applied to protein engineering. A paper published by the University of Wisconsin–Madison demonstrated that a ‘self-driving’ laboratory comprising robotic equipment directed by an AI model successfully reengineered enzymes to have enhanced thermal tolerance, without any input from humans.

A fully autonomous system for protein engineering

With enhanced capabilities of AI, such as knowledge and reasoning for experimental design, prediction of protein/chemical properties, and programming capabilities to control automated lab equipment and analyze output data, we see great potential in using AI and robotics to completely automate the early drug discovery process with high throughput and efficiency.

Here are more recent advances related to AI-based drug discovery.

• AI algorithms can be explainable, providing insights into the chemical structures used to discover novel antibiotics.
• Lingo3DMol, an AI model that combines language models and geometric deep learning technology, can generate small molecules with favorable binding in the target pockets.