Bayesian estimation of Pseudomonas aeruginosa viscoelastic properties

Pseudomonas aeruginosa biofilms are relevant for a variety of disease settings including cystic fibrosis infections. Biofilms are initiated by individual bacteria that undergo a phenotypic switch and produce various types of extracellular polymeric substances such as Psl, Pel, and alginate, which are involved in biofilm development. However, the viscoelastic characteristics of such polysaccharides are not fully explored in literature. For this purpose, we developed a mathematical model to study the rheological behavior of components of three biofilms: P. aeruginosa PAO1, isogenic PAO1ΔwspF, and their mucoid variant PAO1mucA22. Our model consists of a combination of springs and dashpots, which represent the elasticity and viscosity of the substances, respectively. Using a Bayesian inference to estimate these viscoelastic properties, we separate the characteristics and share of each polysaccharide in biofilm structure. A Monte Carlo Markov Chain algorithm is used to estimate these properties in the three biofilm variants, which helps us understand the composition of the biofilm at different stages of their development. This understanding may be used for targeted treatments designed to manipulate the biofilm, making it easier to remove.