Computer simulations of biofilm spatial structure

Abstract

Surface attached communities of bacteria, known as biofilms, can grow in a variety of morphological forms. Classifying these distinct spatial structures is important for understanding the link between the spatial structure of biofilms and other characteristics of biofilm behaviour, such as their genetic diversity and evolutionary behaviour. In this thesis, I examine the phase behaviour of biofilm global spatial structure and link it to the genetic spatial structure, or the spatial distribution of the different cell linages within the biofilm.

I use the iDynoMiCS agent-based biofilm modelling software to simulate Pseudomonas aeruginosa biofilms in a flow cell set up. I develop additional simulation methods to allow the iDynoMiCS software to reach long time scales where the growth dynamics have reaches their steady state and the phase behaviour can be examined.

I find three distinct phases of biofilm growth, which can be classified based on the behaviour of the their steady state interface roughness trajectories and the thickness of the layer of actively growing cells at the top of the biofilm, or 'active layer'. I observe that local gaps in the active layer thickness can cause the biofilm interface to become stationary, or ‘pinned’, relative to the moving front. I distinguish three phases based upon this behaviour: a smooth phase in which there are no pinning sites, a depinned phase in which pinning sites arise but close up again and a pinned phase in which pinning sites arise but cannot be overcome. I find that the variations in the active layer are particularly important in influencing the interface roughness, and argue that the normalised standard deviation of the active layer is an appropriate order parameter for the transition I observe.

I go on to investigate the relationship between the three phases of biofilm growth the genetic diversity. I find that the spatial distribution of different cell lineages is distinctive in each of the phases identified because of the behaviour of the interface roughness and the active layer. Specifically, I observe that pinning sites act as points of diversity loss, while peaks of biofilm fingers are sites of diversity gain via division and mutation.