Biofilm and swimmers interaction : processing and analysis of movies

Resistance of micro-organisms to disinfection is frequently associated with the presence of biofilms on surfaces. To partially overcome biofilms resistance to biocides alternative strategies are being developped. So a wealth of biological antibiofilm agents are under development in the research laboratories but there are only few products on the market and none of them proved to be universal antibiofilm molecules. However, most of these antibiofilm agents may benefit from the "swimmers" approach that would act synergistically by allowing a better infiltration and/or an in situ delivery in the targets biofilms. The goal of the project, Greenswimmers 1, is to identify the "best" bacteria for this purpose of infiltration and biocide delivery. In this technical report we explain how we evaluated the efficiency of swimmers with respect to infiltration of Stapyloccocus Aureus biofilms. This evaluation relies on the kinematics properties of swimmers and on the kinematics of the porous network created by the swimmers, in interaction with the biofilm. Practically, we addressed the tracking of the swimmers with the successful multiple particle tracking algorithm, U-Track, we improved in order to exploit the information from two channels of the biofilm video under the microscope. We also combined swimmers detection and pore segmentation for a more robust estimation of pores. A first step in pore segmentation consists in partitioning the images from the movies with superpixel decomposition. Eventually, more than three hundreds movies were analyzed for eight species and several strains of each of them. Parameters such as speed, persistence, pore area, e-DNA proportion were estimated from the data.