BPS2026 - Applying coarse-grained molecular modeling to generate synthetic data for single-molecule optical trapping and analysis

Emerging optofluidic trapping techniques have made possible the measurement of single-biomolecule kinetics, including interaction and translational diffusion processes, without the need for target labeling or surface conjugation. Although previous work has succeeded in identifying differently sized proteins in heterogenous solutions from the magnitude of forward scattered light intensity (a function of protein size), challenges remain in quickly distinguishing similarly sized proteins. Building on previous modeling efforts to predict molecular polarizability and rotational motion from structure, here we detail a biophysical model to explicitly track the position, orientation and directional light scattering of structures from the protein data bank (PDB) interacting with arbitrary electromagnetic fields. Coupling this model to a simulated beam source and multi-site (i.e., quadrant) photodetector, we then present a method to generate synthetic transmission data for optically trapped biomolecules. We further describe a method to analyze this synthetic data for identifying features, including rotational diffusivity, as well as mass and translational diffusivity. Registering these features with associated confidence intervals, we then use mass-action informed hidden Markov models to test the model's ability to discriminate between both multiple proteins and single-protein conformational states over time. This work can be used to increase the accuracy of identifying similarly size proteins in heterogenous solutions and is expected to support future data collection efforts where samples are limited or costly to obtain, such as in applications for personalized medicine and large-scale biological systems modeling.