Statistical parameter inference of bacterial swimming strategies

TL;DR

This paper develops a stochastic model of E.coli swimming, inferring key parameters from experimental data, and reveals how chemotaxis influences tumble behavior and speed fluctuations.

Contribution

It introduces a detailed Langevin-based framework for bacterial swimming, enabling parameter inference and analysis of chemotactic strategies from experimental data.

Findings

Lower tumble rate when swimming up the gradient

Smaller mean tumble angle during chemotaxis

Increased speed fluctuations when moving up the gradient

Abstract

We provide a detailed stochastic description of the swimming motion of an E.coli bacterium in two dimension, where we resolve tumble events in time. For this purpose, we set up two Langevin equations for the orientation angle and speed dynamics. Calculating moments, distribution and autocorrelation functions from both Langevin equations and matching them to the same quantities determined from data recorded in experiments, we infer the swimming parameters of E.coli . They are the tumble rate ${\lambda}$, the tumble time $r^{-1}$ , the swimming speed $v_0$ , the strength of speed fluctuations ${\sigma}$, the relative height of speed jumps ${\eta}$, the thermal value for the rotational diffusion coefficient $D_0$ , and the enhanced rotational diffusivity during tumbling $D_T$ . Conditioning the observables on the swimming direction relative to the gradient of a chemoattractant, we infer the chemotaxis strategies of E.coli . We confirm the classical strategy of a lower tumble rate for swimming up the gradient but also a smaller mean tumble angle (angle bias). The latter is realized by shorter tumbles as well as a slower diffusive reorientation. We also find that speed fluctuations are increased by about 30% when swimming up the gradient compared to the reversed direction.