Tracking bacillus subtilis in 2D and study of its heterogeneous motility with run-and-tumble model

Abstract

Bacteria swim in the medium by consuming energy, and they can be considered as a living active matter. Swimming of a flagellated bacterium is often described by run-and-tumble (RT) dynamics, which models the trajectory of particle as a combination of run and tumble states. The motility of bacteria in the real world is affected by various natural stochasticities and heterogeneities. In this work, we study the motility of bacteria by comparing the trajectory of swimming B. subtilis with that of RT simulation. For the data collection, we adopt instance embedding, which is a deep learning (DL) based instance segmentation method, to detect the positions of bacteria from experimental video. Training data required for DL is prepared by generating synthetic images and trajectories of bacteria based on experimental data. In RT simulation, we introduce heterogeneity by dispersing the internal speed of particles while stochasticity is reflected through diffusion parameters. As a result, our heterogeneity model successfully reproduces the probability distribution of relative displacement of the experimental trajectory, whereas the simulation with homogeneous model results in much less variance in the x directional distribution of the relative displacement.