Joint open-source initiative combines the Merck portfolio of use cases with the Acceleration Consortium’s world-leading excellence in self-driving labs.
The Acceleration Consortium, based at the University of Toronto (U of T), and leading science and technology company Merck have confirmed their AI-driven experimentation planner Bayesian Back End (BayBE) is now available open-source on GitHub – with an unrestricted Apache 2.0 license.
The joint open-source initiative combines the Merck portfolio of use cases with the Acceleration Consortium’s recognised world-leading excellence in self-driving labs.
Alán Aspuru-Guzik, U of T Professor of Chemistry and Computer Science and Director of the Acceleration Consortium, said: “The release and maintenance of open-source code for scientific discovery is imperative for the advancement of self-driving labs. This collaboration moves the needle of multi-stakeholder work in the area of self-driving labs.
I am thrilled about our collaboration with Merck on developing and publishing software for AI-assisted experimental planning.
As society faces ever-growing challenges, we have no time for science as usual. With this software, we can revolutionize the way experiments are designed and conducted, accelerating discoveries and Driving progress in ways we have never imagined before,” said Alán Aspuru-Guzik, Professor of Chemistry and Computer Science at the University of Toronto, and Director of the Acceleration Consortium, which recently launched a seven-year program worth CA$ 200 million, supported by the Canada First Research Excellence Fund.
“This development is a great outcome of our focus on ‘innovation powered by data and digital’. Together with our partners at the Acceleration Consortium, we continue to push productivity with digital tools such as BayBE. Merck continues to invest in digital
technologies that can disrupt the healthcare, life science and electronics industries,” said Laura Matz, Chief Science and Technology Officer at Merck.
“BayBE unites several advanced technologies under one umbrella and focuses on making them useful for industrial purposes. While it already has many internal use cases, we are excited to share it with a wider community through open source. What started out as a cross-sectorial advancement can now become a cross-industrial one,” Matz said.
BayBE was built jointly across all three business sectors of Merck. It is a general-purpose toolbox for smart iterative experimentation with emphasis on important add-ons for chemistry and materials science. It enables a more systematic approach by providing recommendations for the next best experiment, leading to better results faster.
BayBE can also act as the ‘brain’ for automated equipment, enabling entirely closed-loop self-Driving laboratories.
The traditional approach for design of experiments is largely based on intuition and experience of the experimentalist.
This can lead to considerable variation between different labs and is particularly challenging for complex campaigns that aim to optimize numerous properties simultaneously. Merck faces these challenges on an every-day basis, for instance as part of experimental optimization campaigns in research, product development and operations.
Artificial Intelligence (AI) enables novel ways of tackling these problems and reducing the time needed and money spent as well as increasing sustainability.
The BayBE software already powers dozens of use-cases at Merck, for instance:
- VRP ExcipientFinder: Part of the viscosity reduction platform service of the Life Science business sector of Merck; a tool to accelerate selection of viscosity reducing excipients
- BayChem: Self-service experimental planner available to everyone at Merck, directly enabling lab users
- Self-Driving autonomous flow chemistry at Merck: Closed-loop platform in the R&D of the Life Science business sector that optimizes chemical reactions fully autonomously
In 2023, the University of Toronto was awarded a $200-million grant from the Canada First Research Excellence Fund (CFREF) to revolutionize the speed and impact of scientific discovery through the Acceleration Consortium.
The funding – the largest federal research grant ever awarded to a Canadian university – supports the consortium’s work on self-Driving labs that combine Artificial Intelligence, robotics and advanced computing to discover new materials and molecules in a fraction of the usual time and cost.
Applications include everything from life-saving medications and biodegradable plastics to low-carbon cement and renewable energy.
Researchers in the consortium have used the technology to develop a potential cancer Drug in just 30 days – a process that typically takes years or even decades.
The federal government’s critical support of this initiative builds on years of strategic planning and decisions in this space by the University and the federal government, including the 2017 launch of the Pan-Canadian Artificial Intelligence Strategy that helped cement Toronto’s status as a global hub for a revolutionary technology.
Launched as an Institutional Strategic Initiative in 2021, the Acceleration Consortium brings together partners from academia, government and industry to accelerate the discovery of materials and molecules needed for a sustainable future.
The consortium aims to reduce the time and cost of bringing advanced materials to market, from an average of 20 years and $100 million to as little as one year and $1 million.
The CFREF funding, along with additional support from U of T – which includes an investment of $130 million to expand facilities to house the Acceleration Consortium’s state-of-the-art labs at the Lash Miller Chemical Laboratories building on the St. George campus – will help secure the researchers, spaces and partnerships needed to build a world-leading center for accelerated materials discovery and innovation.
The funding will also help the consortium rapidly create high quality datasets to better train AI models and validate the model’s predictions in real time.
That, in turn, will Dramatically accelerate the discovery and development of molecules and materials for a wide range of industries.
Overall, the Acceleration Consortium comprises nearly 100 researchers across a wide variety of disciplines, including AI, computer science, mathematics, chemistry, economics, engineering, materials science, mechatronics, biology, pharmacology, robotics, technoscience and more.
It also includes 30 partners from the private and public sector, including the University of British Columbia – a lead partner on the grant.
