CDS&E: Accelerating the Computational Design of Nanoparticle-Protein Interactions via Molecular Simulations, Topological Data Analysis, and Machine Learning

Cells secrete nano-scale sacs that contain signaling molecules, enzymes, and sometimes RNA. These sacs, referred to as vesicles, have been repurposed for drug delivery or to carry sensing molecules. Synthetic nanoparticles (NPs) are replacing these vesicles. They exhibit higher uniformity. The main objective of this project is to develop computational methods to design NPs for use with a variety of cell types. In addition, students at UW-Madison and the University of Puerto Rico - Mayagüez will participate in a new course on the intersection of data science and machine learning. Undergraduate research experiences will be offered to students of both institutions as well. Computational tools to predict the binding of proteins to functionalized NPs will be developed. Specifically, all-atom molecular dynamics (MD) simulations and topological data analysis will be combined to compute topological features that capture generalizable chemical and geometric properties of both NP and protein surfaces. This approach should overcome limitations of current protein-protein interaction models that utilize protein-specific features and omit consideration of conformational fluctuations or explicit solvent interactions. Machine learning models will then be trained using these surface features to predict specific NP-protein binding sites. The technical approach will be developed in three major tasks. Task 1 will determine surface representations and topological descriptors for predicting protein-protein and NP-biosystem interactions. Task 2 will use these data representations to train models for predicting NP-protein interactions. Task 3 will demonstrate the design of NPs that (1) selectively bind target proteins, (2) bind target proteins in a specific orientation, or (3) bind a range of common proteins. The outcome of this project will be new and scalable computational tools for designing NPs with targeted protein interactions to enable diverse biomedical applications, as well as new mechanistic insight into the nanoscale factors impacting NP-protein binding. 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.