A model-based approach was developed to detect interspecific interactions during biofilm development. This approach relied on the comparison of experimental data with a simple null model of biofilm growth dynamics where individual species grew independently of one another, except that they competed for space. Such a model was directly parameterized with a 4D confocal image series of biofilms and then used as a null model to detect interspecific interactions between pairs of bacterial species. This approach was tested in two bispecific competitive trials. In the first trial, the progressive exclusion of Pseudomonas fluorescens by Pseudomonas putida appeared to be due solely to the different intrinsic growth rates of the two strains. In contrast, modelling results suggested the presence of interference competition between Pseudomonas aeruginosa and P. putida in mixed biofilms. The authors' approach enables the detection of ecologically relevant interactions which constitute a prerequisite to building a comprehensive view of the dynamics and functioning of spatially structured bacterial communities.
Keywords: Pseudomonas spp.; approximate Bayesian computation; biofilm; interference competition; interspecific interaction; spatial statistics; time-lapse confocal microscopy.