IADR Abstract Archives
Computational Assays for the Characterization of Undisrupted Biofilms' Pathogenicity
Objectives: One of the hallmarks of bacterial biofilms is their reduced sensitivity to antimicrobial agents – in essence, biofilms provide a “safe haven” to bacteria by facilitating attachment to surfaces, limiting diffusion and providing protection from surface wetting - thus preventing efficient interaction between antibiotics and bacterial cells embedded within the biofilm. Accurately measuring and comparing the effects of anti-biofilm agents is crucial in biofilm research, in the clinic or lab.
In this work, we provide a computational framework for the assessment of structural factors that contribute to biofilm pathogenicity in the oral cavity. We present a comprehensive cross-species software that provides a series of quantitative computational ‘assays’ that together form a model for biofilm pathogenicity. We demonstrate this tool on the biofilm-forming bacterium Streptococcus mutans, widely regarded as a major contributor to the development of dental caries.
Methods: Confocal images of S. mutans biofilms with markers for live/dead and EPS, were analyzed in MATLAB software using mathematical, image processing and machine learning tools.
Results: Our software offers support for several computational ‘assays’ that correlate with biofilm pathogenicity – examples include biofilm porosity, surface-specific characteristics and identification of distinct areas in biofilms based on classification algorithms that take into account individual pixel context and its surroundings, thus representing bacterial cells in a true biofilm environment. Furthermore, we present an accurate comparison of biofilms (e.g., treated vs. untreated) using depth-dependent histograms that can pinpoint and differentiate the effects of a particular treatment on specific areas or cellular context, as opposed to a global effect.
Conclusions: Confocal images are ubiquitous in the study of biofilms – indeed, they are frequently used to visualize effects of various antimicrobial treatments. Here we introduce a standardized software tool that can be used to accurately characterize and compare undisrupted biofilms, thus providing a ‘first of its kind’ model for structural pathogenicity.
In this work, we provide a computational framework for the assessment of structural factors that contribute to biofilm pathogenicity in the oral cavity. We present a comprehensive cross-species software that provides a series of quantitative computational ‘assays’ that together form a model for biofilm pathogenicity. We demonstrate this tool on the biofilm-forming bacterium Streptococcus mutans, widely regarded as a major contributor to the development of dental caries.
Methods: Confocal images of S. mutans biofilms with markers for live/dead and EPS, were analyzed in MATLAB software using mathematical, image processing and machine learning tools.
Results: Our software offers support for several computational ‘assays’ that correlate with biofilm pathogenicity – examples include biofilm porosity, surface-specific characteristics and identification of distinct areas in biofilms based on classification algorithms that take into account individual pixel context and its surroundings, thus representing bacterial cells in a true biofilm environment. Furthermore, we present an accurate comparison of biofilms (e.g., treated vs. untreated) using depth-dependent histograms that can pinpoint and differentiate the effects of a particular treatment on specific areas or cellular context, as opposed to a global effect.
Conclusions: Confocal images are ubiquitous in the study of biofilms – indeed, they are frequently used to visualize effects of various antimicrobial treatments. Here we introduce a standardized software tool that can be used to accurately characterize and compare undisrupted biofilms, thus providing a ‘first of its kind’ model for structural pathogenicity.