Salmonella Motility Simulation

An agent-based, standalone 2D single-cell motility simulation for flagellated bacteria.

This project contains Python scripts to simulate single-cell movement in various predefined conditions. These conditions are currently:

Strains:

• PproA
• WT
• PproB

Environments:

• liquid
• agarose

The project uses predefined parameters obtained from single cell measurements, see data/motility_summary_parameters.csv. An arbitrary number of phenotype x environment combinations can be added by extending the input CSV file. Note: This project does not infer or refit parameters from raw trajectory data.

Contents

• docs/: Rendered web page with motility simulations.
• src/salmonella_motility_simulation/: CLI entry point and simulation implementation, including classes, simulations, and plotting.
• data/motility_summary_parameters.csv: Strain-specific movement parameters.
• data/config.yml: Global movement and obstacle parameters.
• output/: Figures and tables as output from the simulation.
• pyproject.toml: Project configuration and dependencies.

Usage

We use Pixi to manage dependencies and define standard tasks for simulations. If you have Pixi installed, you can run the full simulation with:

pixi run simulation

To run specific functions or scripts, check out the enclosed environments and execute python scripts directly, e.g.:

pixi shell
python -m salmonella_motility_simulation --output output/ppro

Scientific purpose

The goal of this script is to provide a model that is:

• more biologically interpretable than a pure active Brownian particle model,
• much simpler than a large multi-state transport engine,
• directly grounded in observed or pre-estimated summary motility parameters,
• easy to read, explain, and modify.

Model overview

The model simulates the typical movement of flagellated bacteria in terms of subsequent pauses / tumbles / reorientation events:

1. Motile cells in run mode

Motile cells move with:

• a phenotype- and medium-specific run speed,
• gradual angular decorrelation through rotational diffusion,
• a stochastic chance of entering a reorientation state.

In this state, the cell performs the visually dominant long displacements in the simulation.

2. Reorientation state

Cells occasionally pause or strongly slow down for a short time.

During this state:

• translational motion is weak,
• the state lasts for a short random duration,
• when the state ends, the heading changes by a random turn angle.

3. Non-motile cells

A phenotype- and medium-specific fraction of cells are permanently non-motile throughout the simulation.

These cells:

• never enter active run mode,
• undergo only weak passive diffusion,
• contribute short, local trajectories.

Liquid vs. agarose environment

The agarose condition adds one explicit environmental interaction:

• the environment contains static, non-overlapping circular obstacles,
• cells are not initialized inside obstacles,
• if a motile cell overlaps an obstacle after a proposed move, it is projected back to the obstacle surface.

After contact, the cell either:

• slides tangentially along the obstacle with reduced displacement, or
• enters a short stalled state with a phenotype-specific probability.

What the model deliberately does not include

This project represents an illustrative but data-grounded track generator. For reasons of simplicity, the model does currently not include:

• chemotaxis,
• cell-cell interactions,
• hydrodynamics,
• wall accumulation,
• phenotype switching,
• source-target transport metrics,
• pore-network reconstruction,
• direct fitting to raw trajectories.

Model parameters

All strain-specific parameters are stored in data/motility_summary_parameters.csv.

• motile_fraction: Fraction of cells that are assigned to the motile subpopulation at the start of the simulation.
• run_speed_um_s: Speed of a motile cell during run mode, in micrometers per second.
• rotational_diffusion_rad2_s: Angular diffusion coefficient controlling how quickly the heading wanders during run mode. Higher values produce more curved and less persistent trajectories.
• reorientation_rate_s: Rate at which a running cell enters the reorientation state.
• reorientation_duration_s: Mean duration of a single reorientation event.
• turn_angle_sd_rad: Standard deviation of the heading change applied when reorientation ends.
• passive_diffusion_um2_s: Weak translational diffusion used for non-motile cells, reorientation motion, and stalled motion.
• stall_probability: In agarose only, probability that an obstacle contact leads to a short stall instead of tangent sliding.
• stall_mean_duration_s: Mean duration of a stall event in agarose.

Authors

• Concept & initial draft: Marc Erhardt, Maria Giralt Zuniga (Humboldt University Berlin, MPUSP)
• Code review, restructuring, editing: Michael Jahn (MPUSP)

The initial draft was created with assistance of a large language model (LLM).

Citation

Coming soon.