Assessing Bacterial Growth and Transport Parameters from Swim Plate Observations

Understanding pathogen motility through confined environments, such as mucosal layers that protect the lung epithelium, provides valuable insight into bacterial infection and growth mechanisms. A bacterial pathogen transport model was used to calculate density profiles representative of bacteria concentration over space and time. Manipulation of the motility coefficient and growth rate can influence the spreading rate exhibited by a bacterial population. However, for a given experimentally observed bacteria distribution, determining the appropriate combination of these two key transport parameters can be ambiguous. For this, we used simple machine learning techniques to identify model output that would allow us to determine if a growth rate and diffusion coefficient of population bacteria can be estimated based on a selected range of our key parameters as deduced from the literature.A library of model bacterial distributions over a range of parameter combinations was first built in Python.Experimental assays of E. coli bacteria in varying compositions of agar and methylcellulose gel were also conducted to provide bacteria density profiles to evaluate model best-fit. From the library we identified the five best-fit continuum models through use of ordinary least squares regression.Our results found that nearly all parameter sets of motility and growth rate were consistent across the searched assay times and values obtained from the observed swim plates matched to models near the lowest range of growth and diffusion parameter values.

The parameter values for bacterial motility generated from our library were also within a magnitude of +/- 50 µm²/s when compared to direct measurements of single cell motility, providing a more targeted range to more efficiently evaluate these parameters from experimental observations and thus develop more accurate models of bacteria growth in various non-chemotactic environments.