Modeling Cell Motility in Acoustic Pressure Fields using Inference Dynamics

Machine learning is frequently being utilized to model biophysical systems, and many algorithms have been developed to deal with the great variety of biological systems. These algorithms are critical to advances in research and innovation as they can incorporate each system's complexity as well as differences in scale between the cellular and sub-cellular level. However, it has proven difficult to determine the specific dynamics that are at play with respect to determinations of both the nature and magnitude of the mechanical forces that cells are able to apply via their cytoskeletal structures in the presence of macroscopic mechanical stresses. This is complicated by the fact that the underlying motor proteins that drive macroscopically observed cell motility operate at a different time and length scales from the overall motion. Our work applies the Underdamped Langevin Inference (ULI) algorithm to cellular motility in the presence of acoustic pressure in vitro. The ULI algorithm is optimized to determine physical quantities for cells experiencing underdamped Brownian motion. The underdamped regime is a good approximation to the external environment of cells whose motion is driven by underlying cytoskeletal forces. Modeling some of these key motor proteins was the topic of our previous work, in which we looked at the both kinesin-1 and tubulin dimers using the overdamped Stochastic Force Inference (SFI). Similar to the SFI algorithm, ULI takes in either simulated or experimental cell motility data and determines the diffusion and force fields that the cell is experiencing in its aqueous environment. We relate the mechanical forces to the chemical degrees of freedom in the microtubule dynamics driving the cell motion. We model cell response to a variety of different external pressures (square pulses, sinusoidal, continuous, etc.) and compare this to motility data in living cells that have been subjected to ultrasonic acoustic pressure fields. Next steps include in vitro studies of ultrasonic acoustic pressure on moss cells cytoskeletal rearrangement in culture, to validate the algorithm. Future work includes investigations of cytoskeletal mechanics, due to various ultrasonic acoustic pressure fields, for a broad range of eukaryotic cells with applications for low gravity environments and solid tumor metastasis.