NSF-BSF: Predictive modeling of contextual, combinatorial signaling

Nearly every process in the body requires active communication between cells. Cells send and receive molecular signals to one another. These signals control how cells grow, specialize their functions, or die. Mutations that disrupt these communication systems can lead to cancer and other diseases. Similar communication systems function across a wide variety of cell types, organs, and processes. This project tackles the question of how a small set of signaling molecules provide enough precision to activate exactly the right cells, at the right time, in the right place. The projects investigates whether a “combinatorial code” provides this specificity. In this view, different combinations of signaling proteins communicate with different cell types based on the combinations of receptors present on the cell surface. The data obtained from measuring responses between different sets of proteins signals and receptors will be used to train mathematical models. The models will then be tested by programming cells to recognize different combinations of signals. The ability to program how cells receive and interpret signals would be a foundational capability for synthetic biology. It would have applications in tissue engineering, regenerative medicine, and therapeutics. This project will also generate computational tools and educational materials, training the next generation of scientists in quantitative and systems biology, better ensuring a globally competitive workforce. The collaboration between the Weizmann Institute of Science and the California Institute of Technology brings together complementary expertise in experimental biology, mathematical modeling, and synthetic biology. This project advances NSF’s priorities in AI, advanced manufacturing and biotechnology. This project investigates combinatorial signaling in the Bone Morphogenetic Protein (BMP) pathway, which broadly regulates cell identity and tissue maintenance. The pathway employs multiple interacting ligand and receptor variants in a many-to-many architecture that endows individual cells with distinct signal integration properties depending on their receptor expression profile. We will pursue three interrelated aims. First, using a receptor replacement strategy in human cells, we will generate a comprehensive dataset of signaling responses across systematically varied ligand and receptor combinations, including configurations not found in nature, providing rich and unique training data for model development. Second, the project will develop and compare two classes of predictive models, biophysical models, grounded in biochemical parameters, and machine learning models, including random forest and gradient boosting approaches, that offer mechanism agnostic prediction. Third, these validated models will guide the design and experimental realization of synthetic cell types with defined receptor profiles that respond selectively to defined ligand combinations, a capability we term ligand addressing, demonstrating predictive, programmable control over a core signal transduction system. This project advances national priorities in biotechnology, artificial intelligence, and synthetic biology by integrating high throughput measurement with predictive machine learning to enable rational design of cellular signaling. The publicly released dataset, predictive models, and open Python course modules will serve the computational biology community and support undergraduate and graduate education in quantitative and systems biology. This collaborative US/Israel project is supported by the US National Science Foundation and the Israeli Binational Science Foundation. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria. NSF Award ID: 2611606 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Michael Elowitz | Institution: California Institute of Technology, PASADENA, CA | Award Amount: $1,299,997 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2611606 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2611606.html