DL tools have even contributed to the conception of new transformative applications such as image-to-image translation 18 or artificial labelling 19, 20. This was impressively demonstrated for image segmentation 8, 9, 10, 11, object detection and classification 12, 13, quality enhancement and denoising 14, 15, and even prediction of super-resolution images 14, 16, 17 from diffraction-limited images. In recent years, the interest in ML, and particularly deep learning (DL), for bioimage analysis has increased significantly, as their high versatility allows them to perform many different image analysis tasks with high performance and speed 4, 5, 6, 7. In bioimage analysis, ML for example contributed to a better understanding of viral organisation 2 and the mode of action of antimicrobial compounds 3.
Therefore, manual analysis is increasingly replaced by automated analysis, particularly with machine learning (ML) 1. The amount of data collected in microbial studies constantly increases with technical developments, which can become challenging for classical data analysis and interpretation, requiring more complex computational approaches to extract relevant features from the data landscape. It covers large spatial scales ranging from single molecules over individual cells to entire ecosystems. The study of microorganisms and microbial communities is a multidisciplinary approach bringing together molecular biology, biochemistry, and biophysics. We hope this lays a fertile ground for the efficient application of DL in microbiology and fosters the creation of tools for bacterial cell biology and antibiotic research.
Our purposefully-built database of training and testing data aids in novice users’ training, enabling them to quickly explore how to analyse their data through DL. Finally, artificial labelling of cell membranes and predictions of super-resolution images allow for accurate mapping of cell shape and intracellular targets. To also demonstrate the ability of DL to enhance low-phototoxicity live-cell microscopy, we showcase how image denoising can allow researchers to attain high-fidelity data in faster and longer imaging. We showcase different deep learning (DL) approaches for segmenting bright field and fluorescence images of different bacterial species, use object detection to classify different growth stages in time-lapse imaging data, and carry out DL-assisted phenotypic profiling of antibiotic-treated cells. We generated a database of image datasets used to train networks for various image analysis tasks and present strategies for data acquisition and curation, as well as model training. This work demonstrates and guides how to use a range of state-of-the-art artificial neural-networks to analyse bacterial microscopy images using the recently developed ZeroCostDL4Mic platform.