Single-strain behavior predicts responses to environmental pH and osmolality in the gut microbiota.

Changes to gut environmental factors such as pH and osmolality due to disease or drugs correlate with major shifts in microbiome composition; however, we currently cannot predict which species can tolerate such changes or how the community will be affected. Here, we assessed the growth of 92 representative human gut bacterial strains spanning 28 families across multiple pH values and osmolalities in vitro. The ability to grow in extreme pH or osmolality conditions correlated with the availability of known stress response genes in many cases, but not all, indicating that novel pathways may participate in protecting against acid or osmotic stresses. Machine learning analysis uncovered genes or subsystems that are predictive of differential tolerance in either acid or osmotic stress. For osmotic stress, we corroborated the increased abundance of these genes in vivo during osmotic perturbation. The growth of specific taxa in limiting conditions in isolation in vitro correlated with survival in complex communities in vitro and in an in vivo mouse model of diet-induced intestinal acidification. Our data show that in vitro stress tolerance results are generalizable and that physical parameters may supersede interspecies interactions in determining the relative abundance of community members. This study provides insight into the ability of the microbiota to respond to common perturbations that may be encountered in the gut and provides a list of genes that correlate with increased ability to survive in these conditions. IMPORTANCE To achieve greater predictability in microbiota studies, it is crucial to consider physical environmental factors such as pH and particle concentration, as they play a pivotal role in influencing bacterial function and survival. For example, pH is significantly altered in various diseases, including cancers, inflammatory bowel disease, as well in the case of over-the-counter drug use. Additionally, conditions like malabsorption can affect particle concentration. In our study, we investigate how changes in environmental pH and osmolality can serve as predictive indicators of bacterial growth and abundance. Our research provides a comprehensive resource for anticipating shifts in microbial composition and gene abundance during complex perturbations. Moreover, our findings underscore the significance of the physical environment as a major driver of bacterial composition. Finally, this work emphasizes the necessity of incorporating physical measurements into animal and clinical studies to better understand the factors influencing shifts in microbiota abundance.

Deep Learning Detection of Active Pulmonary Tuberculosis at Chest Radiography Matched the Clinical Performance of Radiologists.

Background The World Health Organization (WHO) recommends chest radiography to facilitate tuberculosis (TB) screening. However, chest radiograph interpretation expertise remains limited in many regions. Purpose To develop a deep learning system (DLS) to detect active pulmonary TB on chest radiographs and compare its performance to that of radiologists. Materials and Methods A DLS was trained and tested using retrospective chest radiographs (acquired between 1996 and 2020) from 10 countries. To improve generalization, large-scale chest radiograph pretraining, attention pooling, and semisupervised learning ("noisy-student") were incorporated. The DLS was evaluated in a four-country test set (China, India, the United States, and Zambia) and in a mining population in South Africa, with positive TB confirmed with microbiological tests or nucleic acid amplification testing (NAAT). The performance of the DLS was compared with that of 14 radiologists. The authors studied the efficacy of the DLS compared with that of nine radiologists using the Obuchowski-Rockette-Hillis procedure. Given WHO targets of 90% sensitivity and 70% specificity, the operating point of the DLS (0.45) was prespecified to favor sensitivity. Results A total of 165 754 images in 22 284 subjects (mean age, 45 years; 21% female) were used for model development and testing. In the four-country test set (1236 subjects, 17% with active TB), the receiver operating characteristic (ROC) curve of the DLS was higher than those for all nine India-based radiologists, with an area under the ROC curve of 0.89 (95% CI: 0.87, 0.91). Compared with these radiologists, at the prespecified operating point, the DLS sensitivity was higher (88% vs 75%, P < .001) and specificity was noninferior (79% vs 84%, P = .004). Trends were similar within other patient subgroups, in the South Africa data set, and across various TB-specific chest radiograph findings. In simulations, the use of the DLS to identify likely TB-positive chest radiographs for NAAT confirmation reduced the cost by 40%-80% per TB-positive patient detected. Conclusion A deep learning method was found to be noninferior to radiologists for the determination of active tuberculosis on digital chest radiographs. © RSNA, 2022 Online supplemental material is available for this article. See also the editorial by van Ginneken in this issue.