AI-driven biomanufacturing could revolutionize life-saving drug production, making new treatments easier and faster to produce.
Many modern lifesaving treatments for severe or chronic diseases — such as cancer, metabolic disorders or infectious diseases — depend on biopharmaceuticals, or medicines derived from biological sources.
These treatments, made using living cells, tissues or biological substances like proteins or nucleic acids, often provide greater effectiveness and fewer side effects, says Wei Xie, an assistant professor of mechanical and industrial engineering at Northeastern University.
However, manufacturing biopharmaceuticals is highly complex, Xie says, and fraught with uncertainty. It involves hundreds of biological, physical and chemical factors constantly changing and interacting at molecular, cellular and macroscopic levels.
In addition, there is often very little data available, and it can be scattered and inconsistent, especially as new treatments like cell and gene therapies become more personalized.
To address these challenges, Xie recently received a prestigious CAREER award from the National Science Foundation. Her research will focus on developing a fundamental methodology for using artificial intelligence designed with knowledge of biological mechanisms to enhance biopharmaceutical manufacturing innovations.
“This project will help make lifesaving biopharmaceuticals more rapidly available by accelerating biomanufacturing systems integration and automation, significantly improving capabilities,” Xie says. “It will focus on mechanism-informed AI, a new frontier that will integrate scientific knowledge, including synthetic biology and biosystems, with intelligence and decision making under uncertainty.”
Xie’s research aims to create fundamental AI for biomanufacturing. It will allow to represent knowledge across different manufacturing systems in a unified way while enabling easy-to-understand AI learning from limited data for underlying bioprocessing mechanisms and optimal control strategies within and across different scales.
Xie and a team of Northeastern scientists have previously demonstrated the potential of AI in large-scale manufacturing of pluripotent stem cells, or cells with a natural ability to develop into specialized brain, blood, bone or muscle cells.
As part of that work, Xie and her team proposed an innovative biological system-of-systems framework, or Bio-SoS. This modular model accounts for metabolic reactions within cells; cell-to-cell interaction, which happens with diffusion of nutrients, metabolic waste and other molecular substances through the cell aggregates; and aggregates interactions with their surrounding environment.
The Bio-SoS model also allows combining and fusing data from various production systems from lab to large-scale manufacturing with different cell-growing strategies and improving predictions.
While Bio-SoS demonstrated AI’s potential in one setting, Xie says, her CAREER-funded research aims to create a more general, fundamental methodology for biomanufacturing — one that works down to the molecular level.
“Each biomolecule (such as proteins or RNAs), composed of atoms with charge, is a complex system and can have different structures,” Xie says. “How does that affect molecular interactions? If the structure changes, how can we leverage this information in the manufacturing process to guide sample-efficient and interpretable learning on fundamental mechanisms and optimal control policies?”
A prime example is mRNA vaccines, she says, which must be adjusted for various virus mutations. Xie’s research explores how to create such adaptability through new AI modeling and optimization methodologies, supporting flexible manufacturing and a fast response to pandemics.
Xie is optimistic that emerging sensing technologies, which improve monitoring bioprocesses at molecular and cellular levels, will provide new insights for AI to decode fundamental biological and physical mechanisms and advance their scientific understanding.
Beyond research, Xie aims to help develop a world-leading biomanufacturing workforce pipeline through both training the current workforce and educating graduate, undergraduate and K-12 students.