Imagine a world where the power of chemistry is used in concert with machine learning to solve problems in health care, materials science or energy research. This world is one step closer to becoming a reality with the newly formed Center for Computer Assisted Synthesis. Carnegie Mellon chemists Olexandr Isayev and Gabe Gomes are members of the center, which is funded through a five-year, $20 million grant from the National Science Foundation (NSF).
The Center for Computer Assisted Synthesis (C-CAS) was named one of seven Phase II National Science Foundation Centers for Chemical Innovation (CCI) in the nation. The center will change the way chemists solve critical, real-world problems by developing methods that use machine learning to accelerate the synthesis of molecules.
"C-CAS is building a future where a chemist designs a new molecule and uses AI to work out how to best make that molecule," noted Sean L. Jones, assistant director for the Mathematical and Physical Sciences Directorate at the NSF. "This will allow industries, including pharmaceuticals, biotechnology, fine chemicals, electronics and more, to make molecules more sustainably and affordably. This is a grand research challenge with huge societal impact."
C-CAS helps chemists focus on which molecules should be made, rather than on how to make them. By reducing the time and resources needed to design and optimize synthetic routes, the tools and protocols developed in C-CAS provide data-driven approaches to make synthetic chemistry more predictable and efficient because less time is spent on trial-and-error approaches. The tools developed by C-CAS are then shared with the research community through open-source clearinghouses.
Carnegie Mellon's Isayev, assistant professor of chemistry, and Gomes, assistant professor of chemistry and chemical engineering, are among 14 C-CAS faculty investigators from eight institutions.
"Within the next five years, I would like to push the frontiers of computational chemistry methods and AI to mimic and then supersede the chemical intuition and decision-making of expert scientists," Isayev said. "The critical need for 'Chemical Intelligence' exemplifies the need for advancement beyond the presently available algorithms in two primary ways: one, elevate machine learning from generating data models toward generating expert inferences and conclusions; and two, enable autonomous reasoning about these outcomes and execute an actionable research plan, e.g., synthesis of a molecule. Our work has pioneered research at the interface between machine learning and quantum mechanics. These methods will allow fast but accurate calculations of reaction mechanisms and outcomes."
"The future of chemistry is digital. The team at C-CAS is at the forefront of merging organic reactions development and understanding with state-of-the-art computational methods. It is a privilege to work with some of my heroes!" said Gomes, who was recently featured as one of 2022’s Talented 12 by Chemical & Engineering News.
The research programs in the Gomes group at CMU reflect many of the ideas of C-CAS.
"Our team leverages computational strategies to tackle problems in chemical reactivity. Chemistry-infused machine learning plays a central role in our efforts for reaction optimization. Nature also inspires us, like our approach to catalyst design by computationally mimicking the idea of 2018 Chemistry Nobel Laureate Frances Arnold's directed evolution. The collaborations within C-CAS will amplify what our community can do when discovering and developing new catalysts, materials, drugs and more," Gomes added.
Carnegie Mellon University is building the world's first cloud lab at a university. At the Carnegie Mellon University Cloud Lab, which is scheduled to open in spring 2023, hundreds of scientific instruments will be remote-controlled through a universal platform and programming language. The CMU Cloud Lab is part of the university's ambitious future of science initiative that will revolutionize science through the application of automation, artificial intelligence, machine learning and data science.
"The CMU Cloud Lab will democratize access to experimental sciences while enabling a digital thinking approach to how we design experiments. Our groups are taking advantage of this decentralized approach to testing ideas from our AI models," Gomes said.
"Given the expertise of our laboratories and unique positioning, we are taking on a challenge to revolutionize how chemical research and education are done, accelerating the pace of discovery as demanded by the global challenges our society faces," Isayev noted.
C-CAS will also contribute to making the United States a leader in science and technology through attracting, educating and training a new generation of data chemists that includes novel opportunities for researchers from underrepresented groups.
C-CAS is a nexus of collaboration, innovation and education whose impact is amplified by an extensive network of academic researchers, companies, nonprofits and other research centers. This provides C-CAS with a unique opportunity to help shape the future of synthetic chemistry and the fields that rely on it such as medicine, materials and energy.
The CCI Program is a highly competitive, two-phase program. Phase I CCIs receive resources to develop the science, management and broader impact components of a major research center dedicated to a transformative idea before requesting Phase II funding.
"With its collaborative network of data scientists, computer scientists and synthetic and computational chemists, C-CAS is poised to introduce potentially paradigm-shifting approaches," said David Berkowitz, director of the NSF Division of Chemistry. "This Center for Chemical Innovation promises to add an important data-driven dimension to synthesis to complement the intuition of the synthetic chemist, while building on the principles of mechanistic understanding, atom economy, symmetry and convergency."
C-CAS is led by Olaf Wiest, professor in the Department of Chemistry and Biochemistry at Notre Dame, and brings together the expertise of data scientists, computational and synthetic chemists to change the field of synthetic chemistry from an intuition-driven to a data-driven science.
In addition to Carnegie Mellon's Isayev and Gomes, C-CAS faculty affiliates include Nitesh Chawla and Xiangliang Zhang at Notre Dame, Connor Coley at MIT, Abby Doyle and Wei Wang at UCLA, Hosea Nelson and Sarah Reisman at CalTech, Rob Paton at Colorado State University, Richmond Sarpong and Dean Toste at University of California, Berkeley and Matthew Sigman at the University of Utah.
An important part of C-CAS is the Data Chemists Network, a group of faculty with distinguished expertise ranging from synthetic chemistry to science communication at institutions that typically do not participate in large research centers. The perspectives and skills of Raychelle Burks at American University, Andre Isaacs at College of the Holy Cross, Nicholas Ball at Pomona College, Jessica Kisunzu at Colorado College and Ampofo Darko at University of Tennessee greatly enhance the reach and capabilities of C-CAS.
C-CAS is funded by NSF through CHE-2202693. For additional information on C-CAS, visit ccas.nd.edu.
Original source can be found here.
Alerts Sign-up