Machine learning identification of Pseudomonas aeruginosa strains from colony image data.

When grown on agar surfaces, microbes can produce distinct multicellular spatial structures called colonies, which contain characteristic sizes, shapes, edges, textures, and degrees of opacity and color. For over one hundred years, researchers have used these morphology cues to classify bacteria and guide more targeted treatment of pathogens. Advances in genome sequencing technology have revolutionized our ability to classify bacterial isolates and while genomic methods are in the ascendancy, morphological characterization of bacterial species has made a resurgence due to increased computing capacities and widespread application of machine learning tools. In this paper, we revisit the topic of colony morphotype on the within-species scale and apply concepts from image processing, computer vision, and deep learning to a dataset of 69 environmental and clinical Pseudomonas aeruginosa strains. We find that colony morphology and complexity under common laboratory conditions is a robust, repeatable phenotype on the level of individual strains, and therefore forms a potential basis for strain classification. We then use a deep convolutional neural network approach with a combination of data augmentation and transfer learning to overcome the typical data starvation problem in biological applications of deep learning. Using a train/validation/test split, our results achieve an average validation accuracy of 92.9% and an average test accuracy of 90.7% for the classification of individual strains. These results indicate that bacterial strains have characteristic visual 'fingerprints' that can serve as the basis of classification on a sub-species level. Our work illustrates the potential of image-based classification of bacterial pathogens and highlights the potential to use similar approaches to predict medically relevant strain characteristics like antibiotic resistance and virulence from colony data.

Bacterial Quorum Sensing Allows Graded and Bimodal Cellular Responses to Variations in Population Density.

Quorum sensing (QS) is a mechanism of cell-cell communication that connects gene expression to environmental conditions (e.g., cell density) in many bacterial species, mediated by diffusible signal molecules. Current functional studies focus on qualitatively distinct QS ON/OFF states. In the context of density sensing, this view led to the adoption of a "quorum" analogy in which populations sense when they are above a sufficient density (i.e., "quorate") to efficiently turn on cooperative behaviors. This framework overlooks the potential for intermediate, graded responses to shifts in the environment. In this study, we tracked QS-regulated protease (lasB) expression and showed that Pseudomonas aeruginosa can deliver a graded behavioral response to fine-scale variation in population density, on both the population and single-cell scales. On the population scale, we saw a graded response to variation in population density (controlled by culture carrying capacity). On the single-cell scale, we saw significant bimodality at higher densities, with separate OFF and ON subpopulations that responded differentially to changes in density: a static OFF population of cells and increasing intensity of expression among the ON population of cells. Together, these results indicate that QS can tune gene expression to graded environmental change, with no critical cell mass or "quorum" at which behavioral responses are activated on either the individual-cell or population scale. In an infection context, our results indicate there is not a hard threshold separating a quorate "attack" mode from a subquorate "stealth" mode. IMPORTANCE Bacteria can be highly social, controlling collective behaviors via cell-cell communication mechanisms known as quorum sensing (QS). QS is now a large research field, yet a basic question remains unanswered: what is the environmental resolution of QS? The notion of a threshold, or "quorum," separating coordinated ON and OFF states is a central dogma in QS, but recent studies have shown heterogeneous responses at a single cell scale. Using Pseudomonas aeruginosa, we showed that populations generate graded responses to environmental variation through shifts in the proportion of cells responding and the intensity of responses. In an infection context, our results indicate that there is not a hard threshold separating a quorate "attack" mode and a subquorate "stealth" mode.

Effects of an urban sanitation intervention on childhood enteric infection and diarrhea in Maputo, Mozambique: A controlled before-and-after trial.

We conducted a controlled before-and-after trial to evaluate the impact of an onsite urban sanitation intervention on the prevalence of enteric infection, soil transmitted helminth re-infection, and diarrhea among children in Maputo, Mozambique. A non-governmental organization replaced existing poor-quality latrines with pour-flush toilets with septic tanks serving household clusters. We enrolled children aged 1-48 months at baseline and measured outcomes before and 12 and 24 months after the intervention, with concurrent measurement among children in a comparable control arm. Despite nearly exclusive use, we found no evidence that intervention affected the prevalence of any measured outcome after 12 or 24 months of exposure. Among children born into study sites after intervention, we observed a reduced prevalence of Trichuris and Shigella infection relative to the same age group at baseline (<2 years old). Protection from birth may be important to reduce exposure to and infection with enteric pathogens in this setting.