AstraZeneca is linking up with DeepMatter, a big data firm focused on achieving reproducibility in chemistry, to help improve the productivity of its automated compound synthesis operations.
DeepMatter’s artificial intelligence-powered DigitalGlassware platform captures real-time data from a multisensor probe placed within the reaction vessel—factors such as temperature, pressure, ultraviolet light levels and more, as well as taking measurements from the ambient environment.
Combined with data on solvents, catalysts and reagents, the system monitors, records and analyzes the individual steps necessary to synthesize pharmaceutical compounds.
"We've been impressed with the automated chemistry platforms developed at AstraZeneca sites for autonomous delivery of new lead series,” DeepMatter CEO Mark Warne said in a statement.
“We see an opportunity to draw together knowledge from the DigitalGlassware platform to enable machine learning and AI technologies to increase the certainty of producing a high quality and choice of candidate drug molecules,” Warne said.
AstraZeneca hopes this collaboration will allow researchers to replay and explore the chemical reactions digitally while using machine learning to uncover methods to increase yields or save time and resources. The drugmaker aims to employ the cloud-based platform as it builds single compounds as well as larger libraries.
"Our goal is to transform drug design using innovative digital technologies in combination with automation and AI,” said Michael Kossenjans, AstraZeneca’s associate director for discovery sciences.
“To get potential new medicines to patients faster, we need to reduce the cycle time for lead identification and optimization, and look forward to working with DeepMatter to assess the potential of DigitalGlassware to help with this," Kossenjans said.
AstraZeneca recently adopted a machine learning platform for modeling potential molecules developed by Schrödinger and began incorporating it into its drug discovery work this past September.
Schrödinger’s platform aims to predict how well a potential compound will bind with target proteins, which AstraZeneca hopes will help reduce the number that need to be synthesized at all before a lead candidate is chosen.
At the time, AstraZeneca’s global chemistry R&D lead Garry Pairaudeau described the steps as part of a “strategic goal” to transform the company’s drug design operations using digital technology.