Omnipose: a high-precision morphology-independent solution for bacterial cell segmentation

Clostridioides difficile binary toxin CDT induces biofilm-like persisting microcolonies

Clinical symptoms of Clostridioides difficile infection (CDI) range from diarrhea to pseudomembranous colitis. A major challenge in managing CDI is the high rate of relapse. Several studies correlate the production of CDT binary toxin by clinical strains of C. difficile with higher relapse rates. Although the mechanism of action of CDT on host cells is known, its exact contribution to CDI is still unclear. To understand the physiological role of CDT during CDI, we established two hypoxic relevant intestinal models, Transwell and Microfluidic Intestine-on-Chip systems. Both were challenged with the epidemic strain UK1 CDT+ and its isogenic CDT- mutant. We report that CDT induces mucin-associated microcolonies that increase C. difficile colonization and display biofilm-like properties by enhancing C. difficile resistance to vancomycin. Importantly, biofilm-like microcolonies were also observed in the cecum and colon of infected mice. Hence, our study shows that CDT induces biofilm-like microcolonies, increasing C. difficile persistence and risk of relapse.

Mitochondrial segmentation and function prediction in live-cell images with deep learning

Mitochondrial morphology and function are intrinsically linked, indicating the opportunity to predict functions by analyzing morphological features in live-cell imaging. Herein, we introduce MoDL, a deep learning algorithm for mitochondrial image segmentation and function prediction. Trained on a dataset of 20,000 manually labeled mitochondria from super-resolution (SR) images, MoDL achieves superior segmentation accuracy, enabling comprehensive morphological analysis. Furthermore, MoDL predicts mitochondrial functions by employing an ensemble learning strategy, powered by an extended training dataset of over 100,000 SR images, each annotated with functional data from biochemical assays. By leveraging this large dataset alongside data fine-tuning and retraining, MoDL demonstrates the ability to precisely predict functions of heterogeneous mitochondria from unseen cell types through small sample size training. Our results highlight the MoDL's potential to significantly impact mitochondrial research and drug discovery, illustrating its utility in exploring the complex relationship between mitochondrial form and function within a wide range of biological contexts.

Deep generative modeling of annotated bacterial biofilm images

Biofilms are critical for understanding environmental processes, developing biotechnology applications, and progressing in medical treatments of various infections. Nowadays, a key limiting factor for biofilm analysis is the difficulty in obtaining large datasets with fully annotated images. This study introduces a versatile approach for creating synthetic datasets of annotated biofilm images with employing deep generative modeling techniques, including VAEs, GANs, diffusion models, and CycleGAN. Synthetic datasets can significantly improve the training of computer vision models for automated biofilm analysis, as demonstrated with the application of Mask R-CNN detection model. The approach represents a key advance in the field of biofilm research, offering a scalable solution for generating high-quality training data and working with different strains of microorganisms at different stages of formation. Terabyte-scale datasets can be easily generated on personal computers. A web application is provided for the on-demand generation of biofilm images.

Multimodal screen identifies noise-regulatory proteins

Gene-expression noise can influence cell-fate choices across pathology and physiology. However, a crucial question persists: do regulatory proteins or pathways exist that control noise independently of mean expression levels? Our integrative approach, combining single-cell RNA sequencing with proteomics and regulator enrichment analysis, identifies 32 putative noise regulators. SON, a nuclear speckle-associated protein, alters transcriptional noise without changing mean expression levels. Furthermore, SON's noise control can propagate to the protein level. Long-read and total RNA sequencing shows that SON's noise control does not significantly change isoform usage or splicing efficiency. Moreover, SON depletion reduces state switching in pluripotent mouse embryonic stem cells and impacts their fate choice during differentiation. Collectively, we demonstrate a class of proteins that control noise orthogonally to mean expression levels. This work serves as a proof of concept that can identify other functional noise regulators throughout development and disease progression.

Differential Responses of Methylobacterium and Sphingomonas Species to Multispecies Interactions in the Phyllosphere

The leaf surface, known as the phylloplane, presents an oligotrophic and heterogeneous environment due to its topography and uneven distribution of resources. Although it is a challenging environment, leaves support abundant bacterial communities that are spatially structured. However, the factors influencing these spatial distribution patterns are not well understood. To study the changes in population density and spatial distribution of bacteria in synthetic communities, the behaviour of two common bacterial groups in the Arabidopsis thaliana leaf microbiota-Methylobacterium (methylobacteria) and Sphingomonas (sphingomonads)-was examined. Using synthetic communities consisting of two or three species, the hypothesis was tested that the presence of a third species affects the density and spatial interaction of the other two species. Results indicated that methylobacteria exhibit greater sensitivity to changes in population densities and spatial patterns, with higher intra-genus competition and lower densities and aggregation compared to sphingomonads. Pairwise comparisons were insufficient to explain the shifts observed in three-species communities, suggesting that higher-order interactions influence the structuring of complex communities. This emphasises the role of multispecies interactions in determining spatial patterns and community dynamics on the phylloplane.

Machine Learning Models for Segmentation and Classification of Cyanobacterial Cells

Timelapse microscopy has recently been employed to study the metabolism and physiology of cyanobacteria at the single-cell level. However, the identification of individual cells in brightfield images remains a significant challenge. Traditional intensity-based segmentation algorithms perform poorly when identifying individual cells in dense colonies due to a lack of contrast between neighboring cells. Here, we describe a newly developed software package called Cypose which uses machine learning (ML) models to solve two specific tasks: segmentation of individual cyanobacterial cells, and classification of cellular phenotypes. The segmentation models are based on the Cellpose framework, while classification is performed using a convolutional neural network named Cyclass. To our knowledge, these are the first developed ML-based models for cyanobacteria segmentation and classification. When compared to other methods, our segmentation models showed improved performance and were able to segment cells with varied morphological phenotypes, as well as differentiate between live and lysed cells. We also found that our models were robust to imaging artifacts, such as dust and cell debris. Additionally, the classification model was able to identify different cellular phenotypes using only images as input. Together, these models improve cell segmentation accuracy and enable high-throughput analysis of dense cyanobacterial colonies and filamentous cyanobacteria.

GNCnn: A QuPath extension for glomerulosclerosis and glomerulonephritis characterization based on deep learning

The digitalization of traditional glass slide microscopy into whole slide images has opened up new opportunities for pathology, such as the application of artificial intelligence techniques. Specialized software is necessary to visualize and analyze these images. One of these applications is QuPath, a popular bioimage analysis tool. This study proposes GNCnn, the first open-source QuPath extension specifically designed for nephropathology. It integrates deep learning models to provide nephropathologists with an accessible, automatic detector and classifier of glomeruli, the basic filtering units of the kidneys. The aim is to offer nephropathologists a freely available application to measure and analyze glomeruli to identify conditions such as glomerulosclerosis and glomerulonephritis. GNCnn offers a user-friendly interface that enables nephropathologists to detect glomeruli with high accuracy (Dice coefficient of 0.807) and categorize them as either sclerotic or non-sclerotic, achieving a balanced accuracy of 98.46%. Furthermore, it facilitates the classification of non-sclerotic glomeruli into 12 commonly diagnosed types of glomerulonephritis, with a top-3 balanced accuracy of 84.41%. GNCnn provides real-time updates of results, which are available at both the glomerulus and slide levels. This allows users to complete a typical analysis task without leaving the main application, QuPath. This tool is the first to integrate the entire workflow for the assessment of glomerulonephritis directly into the nephropathologists' workspace, accelerating and supporting their diagnosis.

Structural and functional analyses of nematode-derived antimicrobial peptides support the occurrence of direct mechanisms of worm-microbiota interactions

The complex relationships between gastrointestinal (GI) nematodes and the host gut microbiota have been implicated in key aspects of helminth disease and infection outcomes. Nevertheless, the direct and indirect mechanisms governing these interactions are, thus far, largely unknown. In this proof-of-concept study, we demonstrate that the excretory-secretory products (ESPs) and extracellular vesicles (EVs) of key GI nematodes contain peptides that, when recombinantly expressed, exert antimicrobial activity in vitro against Bacillus subtilis. In particular, using time-lapse microfluidics microscopy, we demonstrate that exposure of B. subtilis to a recombinant saposin-domain containing peptide from the 'brown stomach worm', Teladorsagia circumcincta, and a metridin-like ShK toxin from the 'barber's pole worm', Haemonchus contortus, results in cell lysis and significantly reduced growth rates. Data from this study support the hypothesis that GI nematodes may modulate the composition of the vertebrate gut microbiota directly via the secretion of antimicrobial peptides, and pave the way for future investigations aimed at deciphering the impact of such changes on the pathophysiology of GI helminth infection and disease.

CellProfiler plugins - An easy image analysis platform integration for containers and Python tools

CellProfiler is a widely used software for creating reproducible, reusable image analysis workflows without needing to code. In addition to the >90 modules that make up the main CellProfiler program, CellProfiler has a plugins system that allows for the creation of new modules which integrate with other Python tools or tools that are packaged in software containers. The CellProfiler-plugins repository contains a number of these CellProfiler modules, especially modules that are experimental and/or dependency-heavy. Here, we present an upgraded CellProfiler-plugins repository, an example of accessing containerised tools, improved documentation and added citation/reference tools to facilitate the use and contribution of the community.

OmniSegger: A time-lapse image analysis pipeline for bacterial cells

Time-lapse microscopy is a powerful tool for studying the cell biology of bacterial cells. The development of pipelines that facilitate the automated analysis of these datasets is a long-standing goal of the field. In this paper, we describe OmniSegger , an updated version of our SuperSegger pipeline, developed as an open-source, modular, and holistic suite of algorithms whose input is raw microscopy images and whose output is a wide range of quantitative cellular analyses, including dynamical cell cytometry data and cellular visualizations. The updated version described in this paper introduces two principal refinements: (i) robustness to cell morphologies and (ii) support for a range of common imaging modalities. To demonstrate robustness to cell morphology, we present an analysis of the proliferation dynamics of Escherichia coli treated with a drug that induces filamentation. To demonstrate extended support for new image modalities, we analyze cells imaged by five distinct modalities: phase-contrast, two brightfield modalities, and cytoplasmic and membrane fluorescence. Together, this pipeline should greatly increase the scope of tractable analyses for bacterial microscopists.

Artificial Intelligence-Assisted Automatic Raman-Activated Cell Sorting (AI-RACS) System for Mining Specific Functional Microorganisms in the Microbiome

The microbiome represents the natural presence of microorganisms, and exploring, understanding, and leveraging its functions will bring about significant breakthroughs in life sciences and applications. Raman-activated cell sorting (RACS) enables the correlation of phenotype and genotype at the single-cell level, offering a solution to the bottleneck in microbial community functional analysis caused by challenges in cultivating diverse microorganisms. However, current labor-intensive manual procedures fall short in catering to the demands of single-cell functional analysis in microbial communities. To address this issue, we developed an artificial intelligence-assisted Raman-activated cell sorting system (AI-RACS) that integrates precise single-cell positioning, automated data collection, optical tweezers capture, and single-cell printing to elevate microbial single-cell RACS from manual to automated, validating the efficacy of the system by isolating aluminum-tolerant microbes from acidic soil microbiota. Leveraging the AI-RACS framework, we sorted 13 strains from red soil samples under near-in situ conditions, with all demonstrating strong aluminum tolerance. AI-RACS efficiently segregates microbial cells from intricate environmental samples, investigating their functional attributes and presenting a novel tool for microbial research and applications.

Determining the rate-limiting processes for cell division in Escherichia coli

A critical cell cycle checkpoint for most bacteria is the onset of constriction when the septal peptidoglycan synthesis starts. According to the current understanding, the arrival of FtsN to midcell triggers this checkpoint in Escherichia coli. Recent structural and in vitro data suggests that recruitment of FtsN to the Z-ring leads to a conformational switch in actin-like FtsA, which links FtsZ protofilaments to the cell membrane and acts as a hub for the late divisome proteins. Here, we investigate this putative pathway using in vivo measurements and stochastic cell cycle modeling at moderately fast growth rates. Quantitatively upregulating protein concentrations and determining the resulting division timings shows that FtsN and FtsA numbers are not rate-limiting for the division in E. coli. However, at higher overexpression levels, they affect divisions: FtsN by accelerating and FtsA by inhibiting them. At the same time, we find that the FtsZ numbers in the cell are one of the rate-limiting factors for cell divisions in E. coli. Altogether, these findings suggest that instead of FtsN, accumulation of FtsZ in the Z-ring is one of the main drivers of the onset of constriction in E. coli at faster growth rates.

Cell Segmentation With Globally Optimized Boundaries (CSGO): A Deep Learning Pipeline for Whole-Cell Segmentation in Hematoxylin-and-Eosin-Stained Tissues

Accurate whole-cell segmentation is essential in various biomedical applications, particularly in studying the tumor microenvironment. Despite advancements in machine learning for nuclei segmentation in hematoxylin and eosin (H&E)-stained images, there remains a need for effective whole-cell segmentation methods. This study aimed to develop a deep learning-based pipeline to automatically segment cells in H&E-stained tissues, thereby advancing the capabilities of pathological image analysis. The Cell Segmentation with Globally Optimized boundaries (CSGO) framework integrates nuclei and membrane segmentation algorithms, followed by postprocessing using an energy-based watershed method. Specifically, we used the You Only Look Once (YOLO) object detection algorithm for nuclei segmentation and U-Net for membrane segmentation. The membrane detection model was trained on a data set of 7 hepatocellular carcinomas and 11 normal liver tissue patches. The cell segmentation performance was extensively evaluated on 5 external data sets, including liver, lung, and oral disease cases. CSGO demonstrated superior performance over the state-of-the-art method Cellpose, achieving higher F1 scores ranging from 0.37 to 0.53 at an intersection over union threshold of 0.5 in 4 of the 5 external datasets, compared to that of Cellpose from 0.21 to 0.36. These results underscore the robustness and accuracy of our approach in various tissue types. A web-based application is available at https://ai.swmed.edu/projects/csgo, providing a user-friendly platform for researchers to apply our method to their own data sets. Our method exhibits remarkable versatility in whole-cell segmentation across diverse cancer subtypes, serving as an accurate and reliable tool to facilitate tumor microenvironment studies. The advancements presented in this study have the potential to significantly enhance the precision and efficiency of pathologic image analysis, contributing to better understanding and treatment of cancer.

Spontaneous breaking of symmetry in overlapping cell instance segmentation using diffusion models

Instance segmentation is the task of assigning unique identifiers to individual objects in images. Solving this task requires breaking the inherent symmetry that semantically similar objects must result in distinct outputs. Deep learning algorithms bypass this break-of-symmetry by training specialized predictors or by utilizing intermediate label representations. However, many of these approaches break down when faced with overlapping labels that are ubiquitous in biomedical imaging, for instance for segmenting cell layers. Here, we discuss the reason for this failure and offer a novel approach for instance segmentation based on diffusion models that breaks this symmetry spontaneously. Our method outputs pixel-level instance segmentations matching the performance of models such as cellpose on the cellpose fluorescent cell dataset, while also permitting overlapping labels.

Three-dimensional localization and tracking of chromosomal loci throughout the Escherichia coli cell cycle

The intracellular position of genes may impact their expression, but it has not been possible to accurately measure the 3D position of chromosomal loci. In 2D, loci can be tracked using arrays of DNA-binding sites for transcription factors (TFs) fused with fluorescent proteins. However, the same 2D data can result from different 3D trajectories. Here, we have developed a deep learning method for super-resolved astigmatism-based 3D localization of chromosomal loci in live E. coli cells which enables a precision better than 61 nm at a signal-to-background ratio of ~4 on a heterogeneous cell background. Determining the spatial localization of chromosomal loci, we find that some loci are at the periphery of the nucleoid for large parts of the cell cycle. Analyses of individual trajectories reveal that these loci are subdiffusive both longitudinally (x) and radially (r), but that individual loci explore the full radial width on a minute time scale.

Molecular mechanism of a triazole-containing inhibitor of Mycobacterium tuberculosis DNA gyrase

Antimicrobial resistance remains a persistent and pressing public health concern. Here, we describe the synthesis of original triazole-containing inhibitors targeting the DNA gyrase, a well-validated drug target for developing new antibiotics. Our compounds demonstrate potent antibacterial activity against various pathogenic bacteria, with notable potency against Mycobacterium tuberculosis (Mtb). Moreover, one hit, compound 10a, named BDM71403, was shown to be more potent in Mtb than the NBTI of reference, gepotidacin. Mechanistic enzymology assays reveal a competitive interaction of BDM71403 with fluoroquinolones within the Mtb gyrase cleavage core. High-resolution cryo-electron microscopy structural analysis provides detailed insights into the ternary complex formed by the Mtb gyrase, double-stranded DNA, and either BDM71403 or gepotidacin, providing a rational framework to understand the superior in vitro efficacy on Mtb. This study highlights the potential of triazole-based scaffolds as promising gyrase inhibitors, offering new avenues for drug development in the fight against antimicrobial resistance.

Human formin FHOD3-mediated actin elongation is required for sarcomere integrity in cardiomyocytes

Contractility and cell motility depend on accurately controlled assembly of the actin cytoskeleton. Formins are a large group of actin assembly proteins that nucleate new actin filaments and act as elongation factors. Some formins may cap filaments, instead of elongating them, and others are known to sever or bundle filaments. The Formin HOmology Domain-containing protein (FHOD)-family of formins is critical to the formation of the fundamental contractile unit in muscle, the sarcomere. Specifically, mammalian FHOD3L plays an essential role in cardiomyocytes. Despite our knowledge of FHOD3L's importance in cardiomyocytes, its biochemical and cellular activities remain poorly understood. It has been proposed that FHOD-family formins act by capping and bundling, as opposed to assembling new filaments. Here, we demonstrate that FHOD3L nucleates actin and rapidly but briefly elongates filaments after temporarily pausing elongation, in vitro. We designed function-separating mutants that enabled us to distinguish which biochemical roles are reqùired in the cell. We found that human FHOD3L's elongation activity, but not its nucleation, capping, or bundling activity, is necessary for proper sarcomere formation and contractile function in neonatal rat ventricular myocytes. The results of this work provide new insight into the mechanisms by which formins build specific structures and will contribute to knowledge regarding how cardiomyopathies arise from defects in sarcomere formation and maintenance.

Introducing FISCAS, a Tool for the Effective Generation of Single Cell MALDI-MSI Data

We introduce Fluorescence Integrated Single-Cell Analysis Script (FISCAS), which combines fluorescence microscopy with MALDI-MSI to streamline single-cell analysis. FISCAS enables automated selection of tight measurement regions, thereby reducing the acquisition of off-target pixels, and makes use of established algorithms for cell segmentation and coregistration to rapidly compile single-cell spectra. MALDI-compatible staining of membranes, nuclei, and lipid droplets allows the collection of fluorescence data prior to the MALDI-MSI measurement on a timsTOF fleX MALDI-2. Usefulness of the software is demonstrated by the example of THP-1 cells during stimulated differentiation into macrophages at different time points. In this proof-of-principle study, FISCAS was used to automatically generate single-cell mass spectra along with a wide range of morphometric parameters for a total number of roughly 1300 cells collected at 24, 48, and 72 h after the onset of stimulation. Data analysis of the combined morphometric and single-cell mass spectrometry data shows significant molecular heterogeneity within the cell population at each time point, indicating an independent differentiation of each individual cell rather than a synchronized mechanism. Here, the grouping of cells based on their molecular phenotype revealed an overall clearer distinction of the different phases of differentiation into macrophages and delivered an increased number of lipid signals as possible markers compared with traditional bulk analysis. Utilizing the linkage between mass spectrometric data and fluorescence microscopy confirmed the expected positive correlation between lipid droplet staining and the overall signal for triacylglyceride (TG), demonstrating the usefulness of this multimodal approach.

Coupling of cell growth modulation to asymmetric division and cell cycle regulation in Caulobacter crescentus

In proliferating bacteria, growth rate is often assumed to be similar between daughter cells. However, most of our knowledge of cell growth derives from studies on symmetrically dividing bacteria. In many α-proteobacteria, asymmetric division is a normal part of the life cycle, with each division producing daughter cells with different sizes and fates. Here, we demonstrate that the functionally distinct swarmer and stalked daughter cells produced by the model α-proteobacterium Caulobacter crescentus can have different average growth rates under nutrient-replete conditions despite sharing an identical genome and environment. The discrepancy in growth rate is due to a growth slowdown associated with the cell cycle stage preceding DNA replication (the G1 phase), which initiates in the late predivisional mother cell before daughter cell separation. Both progenies experience a G1-associated growth slowdown, but the effect is more severe in swarmer cells because they have a longer G1 phase. Activity of SpoT, which produces the (p)ppGpp alarmone and extends the G1 phase, accentuates the cell cycle-dependent growth slowdown. Collectively, our data identify a coupling between cell growth, the G1 phase, and asymmetric division that C. crescentus may exploit for environmental adaptation through SpoT activity. This coupling differentially modulates the growth rate of functionally distinct daughter cells, thereby altering the relative abundance of ecologically important G1-specific traits within the population.

Phage-triggered reverse transcription assembles a toxic repetitive gene from a noncoding RNA

Reverse transcription has frequently been co-opted for cellular functions and in prokaryotes is associated with protection against viral infection, but the underlying mechanisms of defense are generally unknown. Here, we show that in the DRT2 defense system, the reverse transcriptase binds a neighboring pseudoknotted noncoding RNA. Upon bacteriophage infection, a template region of this RNA is reverse transcribed into an array of tandem repeats that reconstitute a promoter and open reading frame, allowing expression of a toxic repetitive protein and an abortive infection response. Biochemical reconstitution of this activity and cryo-electron microscopy provide a molecular basis for repeat synthesis. Gene synthesis from a noncoding RNA is a previously unknown mode of genetic regulation in prokaryotes.

Mapping spatial organization and genetic cell-state regulators to target immune evasion in ovarian cancer

The drivers of immune evasion are not entirely clear, limiting the success of cancer immunotherapies. Here we applied single-cell spatial and perturbational transcriptomics to delineate immune evasion in high-grade serous tubo-ovarian cancer. To this end, we first mapped the spatial organization of high-grade serous tubo-ovarian cancer by profiling more than 2.5 million cells in situ in 130 tumors from 94 patients. This revealed a malignant cell state that reflects tumor genetics and is predictive of T cell and natural killer cell infiltration levels and response to immune checkpoint blockade. We then performed Perturb-seq screens and identified genetic perturbations-including knockout of PTPN1 and ACTR8-that trigger this malignant cell state. Finally, we show that these perturbations, as well as a PTPN1/PTPN2 inhibitor, sensitize ovarian cancer cells to T cell and natural killer cell cytotoxicity, as predicted. This study thus identifies ways to study and target immune evasion by linking genetic variation, cell-state regulators and spatial biology.

Bioinformatic analysis reveals the association between bacterial morphology and antibiotic resistance using light microscopy with deep learning

Although it is well known that the morphology of Gram-negative rods changes on exposure to antibiotics, the morphology of antibiotic-resistant bacteria in the absence of antibiotics has not been widely investigated. Here, we studied the morphologies of 10 antibiotic-resistant strains of Escherichia coli and used bioinformatics tools to classify the resistant cells under light microscopy in the absence of antibiotics. The antibiotic-resistant strains showed differences in morphology from the sensitive parental strain, and the differences were most prominent in the quinolone-and β-lactam-resistant bacteria. A cluster analysis revealed increased proportions of fatter or shorter cells in the antibiotic-resistant strains. A correlation analysis of morphological features and gene expression suggested that genes related to energy metabolism and antibiotic resistance were highly correlated with the morphological characteristics of the resistant strains. Our newly proposed deep learning method for single-cell classification achieved a high level of performance in classifying quinolone-and β-lactam-resistant strains.

Cryo-EM structure and polar assembly of the PS2 S-layer of Corynebacterium glutamicum

The polar-growing Corynebacteriales have a complex cell envelope architecture characterized by the presence of a specialized outer membrane composed of mycolic acids. In some Corynebacteriales, this mycomembrane is further supported by a proteinaceous surface layer or 'S-layer', whose function, structure and mode of assembly remain largely enigmatic. Here, we isolated ex vivo PS2 S-layers from the industrially important Corynebacterium glutamicum and determined its atomic structure by 3D cryoEM reconstruction. PS2 monomers consist of a six-helix bundle 'core', a three-helix bundle 'arm', and a C-terminal transmembrane (TM) helix. The PS2 core oligomerizes into hexameric units anchored in the mycomembrane by a channel-like coiled-coil of the TM helices. The PS2 arms mediate trimeric lattice contacts, crystallizing the hexameric units into an intricate semipermeable lattice. Using pulse-chase live cell imaging, we show that the PS2 lattice is incorporated at the poles, coincident with the actinobacterial elongasome. Finally, phylogenetic analysis shows a paraphyletic distribution and dispersed chromosomal location of PS2 in Corynebacteriales as a result of multiple recombination events and losses. These findings expand our understanding of S-layer biology and enable applications of membrane-supported self-assembling bioengineered materials.

Quantitative microbiology with widefield microscopy: navigating optical artefacts for accurate interpretations

Time-resolved live-cell imaging using widefield microscopy is instrumental in quantitative microbiology research. It allows researchers to track and measure the size, shape, and content of individual microbial cells over time. However, the small size of microbial cells poses a significant challenge in interpreting image data, as their dimensions approache that of the microscope's depth of field, and they begin to experience significant diffraction effects. As a result, 2D widefield images of microbial cells contain projected 3D information, blurred by the 3D point spread function. In this study, we employed simulations and targeted experiments to investigate the impact of diffraction and projection on our ability to quantify the size and content of microbial cells from 2D microscopic images. This study points to some new and often unconsidered artefacts resulting from the interplay of projection and diffraction effects, within the context of quantitative microbiology. These artefacts introduce substantial errors and biases in size, fluorescence quantification, and even single-molecule counting, making the elimination of these errors a complex task. Awareness of these artefacts is crucial for designing strategies to accurately interpret micrographs of microbes. To address this, we present new experimental designs and machine learning-based analysis methods that account for these effects, resulting in accurate quantification of microbiological processes.

Protein overabundance is driven by growth robustness

Protein expression levels optimize cell fitness: Too low an expression level of essential proteins will slow growth by compromising essential processes; whereas overexpression slows growth by increasing the metabolic load. This trade-off naïvely predicts that cells maximize their fitness by sufficiency, expressing just enough of each essential protein for function. We test this prediction in the naturally-competent bacterium Acinetobacter baylyi by characterizing the proliferation dynamics of essential-gene knockouts at a single-cell scale (by imaging) as well as at a genome-wide scale (by TFNseq). In these experiments, cells proliferate for multiple generations as target protein levels are diluted from their endogenous levels. This approach facilitates a proteome-scale analysis of protein overabundance. As predicted by the Robustness-Load Trade-Off (RLTO) model, we find that roughly 70% of essential proteins are overabundant and that overabundance increases as the expression level decreases, the signature prediction of the model. These results reveal that robustness plays a fundamental role in determining the expression levels of essential genes and that overabundance is a key mechanism for ensuring robust growth.

Protein overabundance is driven by growth robustness

Protein expression levels optimize cell fitness: Too low an expression level of essential proteins will slow growth by compromising essential processes; whereas overexpression slows growth by increasing the metabolic load. This trade-off naïvely predicts that cells maximize their fitness by sufficiency, expressing just enough of each essential protein for function. We test this prediction in the naturally-competent bacterium Acinetobacter baylyi by characterizing the proliferation dynamics of essential-gene knockouts at a single-cell scale (by imaging) as well as at a genome-wide scale (by TFNseq). In these experiments, cells proliferate for multiple generations as target protein levels are diluted from their endogenous levels. This approach facilitates a proteome-scale analysis of protein overabundance. As predicted by the Robustness-Load Trade-Off (RLTO) model, we find that roughly 70% of essential proteins are overabundant and that overabundance increases as the expression level decreases, the signature prediction of the model. These results reveal that robustness plays a fundamental role in determining the expression levels of essential genes and that overabundance is a key mechanism for ensuring robust growth.

A benchmarked comparison of software packages for time-lapse image processing of monolayer bacterial population dynamics

Time-lapse microscopy offers a powerful approach for analyzing cellular activity. In particular, this technique is valuable for assessing the behavior of bacterial populations, which can exhibit growth and intercellular interactions in a monolayer. Such time-lapse imaging typically generates large quantities of data, limiting the options for manual investigation. Several image-processing software packages have been developed to facilitate analysis. It can thus be a challenge to identify the software package best suited to a particular research goal. Here, we compare four software packages that support the analysis of 2D time-lapse images of cellular populations: CellProfiler, SuperSegger-Omnipose, DeLTA, and FAST. We compare their performance against benchmarked results on time-lapse observations of Escherichia coli populations. Performance varies across the packages, with each of the four outperforming the others in at least one aspect of the analysis. Not surprisingly, the packages that have been in development for longer showed the strongest performance. We found that deep learning-based approaches to object segmentation outperformed traditional approaches, but the opposite was true for frame-to-frame object tracking. We offer these comparisons, together with insight into usability, computational efficiency, and feature availability, as a guide to researchers seeking image-processing solutions.

IMPORTANCE

Time-lapse microscopy provides a detailed window into the world of bacterial behavior. However, the vast amount of data produced by these techniques is difficult to analyze manually. We have analyzed four software tools designed to process such data and compared their performance, using populations of commonly studied bacterial species as our test subjects. Our findings offer a roadmap to scientists, helping them choose the right tool for their research. This comparison bridges a gap between microbiology and computational analysis, streamlining research efforts.

A deterministic, c-di-GMP-dependent program ensures the generation of phenotypically similar, symmetric daughter cells during cytokinesis

Phenotypic heterogeneity in bacteria can result from stochastic processes or deterministic programs. The deterministic programs often involve the versatile second messenger c-di-GMP, and give rise to daughter cells with different c-di-GMP levels by deploying c-di-GMP metabolizing enzymes asymmetrically during cell division. By contrast, less is known about how phenotypic heterogeneity is kept to a minimum. Here, we identify a deterministic c-di-GMP-dependent program that is hardwired into the cell cycle of Myxococcus xanthus to minimize phenotypic heterogeneity and guarantee the formation of phenotypically similar daughter cells during division. Cells lacking the diguanylate cyclase DmxA have an aberrant motility behaviour. DmxA is recruited to the cell division site and its activity is switched on during cytokinesis, resulting in a transient increase in the c-di-GMP concentration. During cytokinesis, this c-di-GMP burst ensures the symmetric incorporation and allocation of structural motility proteins and motility regulators at the new cell poles of the two daughters, thereby generating phenotypically similar daughters with correct motility behaviours. Thus, our findings suggest a general c-di-GMP-dependent mechanism for minimizing phenotypic heterogeneity, and demonstrate that bacteria can ensure the formation of dissimilar or similar daughter cells by deploying c-di-GMP metabolizing enzymes to distinct subcellular locations.

GZMK+CD8+ T cells Target A Specific Acinar Cell Type in Sjögren's Disease

Sjögren's Disease (SjD) is a systemic autoimmune disease without a clear etiology or effective therapy. Utilizing unbiased single-cell and spatial transcriptomics to analyze human minor salivary glands in health and disease we developed a comprehensive understanding of the cellular landscape of healthy salivary glands and how that landscape changes in SjD patients. We identified novel seromucous acinar cell types and identified a population of PRR4+CST3+WFDC2- seromucous acinar cells that are particularly targeted in SjD. Notably, GZMK+CD8 T cells, enriched in SjD, exhibited a cytotoxic phenotype and were physically associated with immune-engaged epithelial cells in disease. These findings shed light on the immune response's impact on transitioning acinar cells with high levels of secretion and explain the loss of this specific cell population in SjD. This study explores the complex interplay of varied cell types in the salivary glands and their role in the pathology of Sjögren's Disease.

Molecular definition of the endogenous Toll-like receptor signalling pathways

Innate immune pattern recognition receptors, such as the Toll-like receptors (TLRs), are key mediators of the immune response to infection and central to our understanding of health and disease1. After microbial detection, these receptors activate inflammatory signal transduction pathways that involve IκB kinases, mitogen-activated protein kinases, ubiquitin ligases and other adaptor proteins. The mechanisms that connect the proteins in the TLR pathways are poorly defined. To delineate TLR pathway activities, we engineered macrophages to enable microscopy and proteomic analysis of the endogenous myddosome constituent MyD88. We found that myddosomes form transient contacts with activated TLRs and that TLR-free myddosomes are dynamic in size, number and composition over the course of 24 h. Analysis using super-resolution microscopy revealed that, within most myddosomes, MyD88 forms barrel-like structures that function as scaffolds for effector protein recruitment. Proteomic analysis demonstrated that myddosomes contain proteins that act at all stages and regulate all effector responses of the TLR pathways, and genetic analysis defined the epistatic relationship between these effector modules. Myddosome assembly was evident in cells infected with Listeria monocytogenes, but these bacteria evaded myddosome assembly and TLR signalling during cell-to-cell spread. On the basis of these findings, we propose that the entire TLR signalling pathway is executed from within the myddosome.

Colony context and size-dependent compensation mechanisms give rise to variations in nuclear growth trajectories

To investigate the fundamental question of how cellular variations arise across spatiotemporal scales in a population of identical healthy cells, we focused on nuclear growth in hiPS cell colonies as a model system. We generated a 3D timelapse dataset of thousands of nuclei over multiple days, and developed open-source tools for image and data analysis and an interactive timelapse viewer for exploring quantitative features of nuclear size and shape. We performed a data-driven analysis of nuclear growth variations across timescales. We found that individual nuclear volume growth trajectories arise from short timescale variations attributable to their spatiotemporal context within the colony. We identified a strikingly time-invariant volume compensation relationship between nuclear growth duration and starting volume across the population. Notably, we discovered that inheritance plays a crucial role in determining these two key nuclear growth features while other growth features are determined by their spatiotemporal context and are not inherited.

Combining machine learning with high-content imaging to infer ciprofloxacin susceptibility in isolates of Salmonella Typhimurium

Antimicrobial resistance (AMR) is a growing public health crisis that requires innovative solutions. Current susceptibility testing approaches limit our ability to rapidly distinguish between antimicrobial-susceptible and -resistant organisms. Salmonella Typhimurium (S. Typhimurium) is an enteric pathogen responsible for severe gastrointestinal illness and invasive disease. Despite widespread resistance, ciprofloxacin remains a common treatment for Salmonella infections, particularly in lower-resource settings, where the drug is given empirically. Here, we exploit high-content imaging to generate deep phenotyping of S. Typhimurium isolates longitudinally exposed to increasing concentrations of ciprofloxacin. We apply machine learning algorithms to the imaging data and demonstrate that individual isolates display distinct growth and morphological characteristics that cluster by time point and susceptibility to ciprofloxacin, which occur independently of ciprofloxacin exposure. Using a further set of S. Typhimurium clinical isolates, we find that machine learning classifiers can accurately predict ciprofloxacin susceptibility without exposure to it or any prior knowledge of resistance phenotype. These results demonstrate the principle of using high-content imaging with machine learning algorithms to predict drug susceptibility of clinical bacterial isolates. This technique may be an important tool in understanding the morphological impact of antimicrobials on the bacterial cell to identify drugs with new modes of action.

The multimodality cell segmentation challenge: toward universal solutions

Cell segmentation is a critical step for quantitative single-cell analysis in microscopy images. Existing cell segmentation methods are often tailored to specific modalities or require manual interventions to specify hyper-parameters in different experimental settings. Here, we present a multimodality cell segmentation benchmark, comprising more than 1,500 labeled images derived from more than 50 diverse biological experiments. The top participants developed a Transformer-based deep-learning algorithm that not only exceeds existing methods but can also be applied to diverse microscopy images across imaging platforms and tissue types without manual parameter adjustments. This benchmark and the improved algorithm offer promising avenues for more accurate and versatile cell analysis in microscopy imaging.

Svetlana a supervised segmentation classifier for Napari

We present Svetlana (SuperVised sEgmenTation cLAssifier for NapAri), an open-source Napari plugin dedicated to the manual or automatic classification of segmentation results. A few recent software tools have made it possible to automatically segment complex 2D and 3D objects such as cells in biology with unrivaled performance. However, the subsequent analysis of the results is oftentimes inaccessible to non-specialists. The Svetlana plugin aims at going one step further, by allowing end-users to label the segmented objects and to pick, train and run arbitrary neural network classifiers. The resulting network can then be used for the quantitative analysis of biophysical phenoma. We showcase its performance through challenging problems in 2D and 3D and provide a comprehensive discussion on its strengths and limits.

Deep Learning Insights into the Dynamic Effects of Photodynamic Therapy on Cancer Cells

Photodynamic therapy (PDT) shows promise in tumor treatment, particularly when combined with nanotechnology. This study examines the impact of deep learning, particularly the Cellpose algorithm, on the comprehension of cancer cell responses to PDT. The Cellpose algorithm enables robust morphological analysis of cancer cells, while logistic growth modelling predicts cellular behavior post-PDT. Rigorous model validation ensures the accuracy of the findings. Cellpose demonstrates significant morphological changes after PDT, affecting cellular proliferation and survival. The reliability of the findings is confirmed by model validation. This deep learning tool enhances our understanding of cancer cell dynamics after PDT. Advanced analytical techniques, such as morphological analysis and growth modeling, provide insights into the effects of PDT on hepatocellular carcinoma (HCC) cells, which could potentially improve cancer treatment efficacy. In summary, the research examines the role of deep learning in optimizing PDT parameters to personalize oncology treatment and improve efficacy.

Dynamic microfluidic single-cell screening identifies pheno-tuning compounds to potentiate tuberculosis therapy

Drug-recalcitrant infections are a leading global-health concern. Bacterial cells benefit from phenotypic variation, which can suggest effective antimicrobial strategies. However, probing phenotypic variation entails spatiotemporal analysis of individual cells that is technically challenging, and hard to integrate into drug discovery. In this work, we develop a multi-condition microfluidic platform suitable for imaging two-dimensional growth of bacterial cells during transitions between separate environmental conditions. With this platform, we implement a dynamic single-cell screening for pheno-tuning compounds, which induce a phenotypic change and decrease cell-to-cell variation, aiming to undermine the entire bacterial population and make it more vulnerable to other drugs. We apply this strategy to mycobacteria, as tuberculosis poses a major public-health threat. Our lead compound impairs Mycobacterium tuberculosis via a peculiar mode of action and enhances other anti-tubercular drugs. This work proves that harnessing phenotypic variation represents a successful approach to tackle pathogens that are increasingly difficult to treat.

Burkholderia thailandensis uses a type VI secretion system to lyse protrusions without triggering host cell responses

To spread within a host, intracellular Burkholderia form actin tails to generate membrane protrusions into neighboring host cells and use type VI secretion system-5 (T6SS-5) to induce cell-cell fusions. Here, we show that B. thailandensis also uses T6SS-5 to lyse protrusions to directly spread from cell to cell. Dynamin-2 recruitment to the membrane near a bacterium was followed by a short burst of T6SS-5 activity. This resulted in the polymerization of the actin of the newly invaded host cell and disruption of the protrusion membrane. Most protrusion lysis events were dependent on dynamin activity, caused no cell-cell fusion, and failed to be recognized by galectin-3. T6SS-5 inactivation decreased protrusion lysis but increased galectin-3, LC3, and LAMP1 accumulation in host cells. Our results indicate that B. thailandensis specifically activates T6SS-5 assembly in membrane protrusions to disrupt host cell membranes and spread without alerting cellular responses, such as autophagy.

A general algorithm for consensus 3D cell segmentation from 2D segmented stacks

Cell segmentation is the fundamental task. Only by segmenting, can we define the quantitative spatial unit for collecting measurements to draw biological conclusions. Deep learning has revolutionized 2D cell segmentation, enabling generalized solutions across cell types and imaging modalities. This has been driven by the ease of scaling up image acquisition, annotation and computation. However 3D cell segmentation, which requires dense annotation of 2D slices still poses significant challenges. Labelling every cell in every 2D slice is prohibitive. Moreover it is ambiguous, necessitating cross-referencing with other orthoviews. Lastly, there is limited ability to unambiguously record and visualize 1000's of annotated cells. Here we develop a theory and toolbox, u-Segment3D for 2D-to-3D segmentation, compatible with any 2D segmentation method. Given optimal 2D segmentations, u-Segment3D generates the optimal 3D segmentation without data training, as demonstrated on 11 real life datasets, >70,000 cells, spanning single cells, cell aggregates and tissue.

Segmentation of hyphae and yeast in fungi-infected tissue slice images and its application in analyzing antifungal blue light therapy

Candida albicans is a pathogenic fungus that undergoes morphological transitions between hyphal and yeast forms, adapting to diverse environmental stimuli and exhibiting distinct virulence. Existing research works on antifungal blue light (ABL) therapy have either focused solely on hyphae or neglected to differentiate between morphologies, obscuring potential differential effects. To address this gap, we established a novel dataset of 150 C. albicans-infected mouse skin tissue slice images with meticulously annotated hyphae and yeast. Eleven representative convolutional neural networks were trained and evaluated on this dataset using seven metrics to identify the optimal model for segmenting hyphae and yeast in original high pixel size images. Leveraging the segmentation results, we analyzed the differential impact of blue light on the invasion depth and density of both morphologies within the skin tissue. U-Net-BN outperformed other models in segmentation accuracy, achieving the best overall performance. While both hyphae and yeast exhibited significant reductions in invasion depth and density at the highest ABL dose (180 J/cm2), only yeast was significantly inhibited at the lower dose (135 J/cm2). This novel finding emphasizes the importance of developing more effective treatment strategies for both morphologies.

Streptomyces umbrella toxin particles block hyphal growth of competing species

Streptomyces are a genus of ubiquitous soil bacteria from which the majority of clinically utilized antibiotics derive1. The production of these antibacterial molecules reflects the relentless competition Streptomyces engage in with other bacteria, including other Streptomyces species1,2. Here we show that in addition to small-molecule antibiotics, Streptomyces produce and secrete antibacterial protein complexes that feature a large, degenerate repeat-containing polymorphic toxin protein. A cryo-electron microscopy structure of these particles reveals an extended stalk topped by a ringed crown comprising the toxin repeats scaffolding five lectin-tipped spokes, which led us to name them umbrella particles. Streptomyces coelicolor encodes three umbrella particles with distinct toxin and lectin composition. Notably, supernatant containing these toxins specifically and potently inhibits the growth of select Streptomyces species from among a diverse collection of bacteria screened. For one target, Streptomyces griseus, inhibition relies on a single toxin and that intoxication manifests as rapid cessation of vegetative hyphal growth. Our data show that Streptomyces umbrella particles mediate competition among vegetative mycelia of related species, a function distinct from small-molecule antibiotics, which are produced at the onset of reproductive growth and act broadly3,4. Sequence analyses suggest that this role of umbrella particles extends beyond Streptomyces, as we identified umbrella loci in nearly 1,000 species across Actinobacteria.

Deep learning-driven imaging of cell division and cell growth across an entire eukaryotic life cycle

The life cycle of biomedical and agriculturally relevant eukaryotic microorganisms involves complex transitions between proliferative and non-proliferative states such as dormancy, mating, meiosis, and cell division. New drugs, pesticides, and vaccines can be created by targeting specific life cycle stages of parasites and pathogens. However, defining the structure of a microbial life cycle often relies on partial observations that are theoretically assembled in an ideal life cycle path. To create a more quantitative approach to studying complete eukaryotic life cycles, we generated a deep learning-driven imaging framework to track microorganisms across sexually reproducing generations. Our approach combines microfluidic culturing, life cycle stage-specific segmentation of microscopy images using convolutional neural networks, and a novel cell tracking algorithm, FIEST, based on enhancing the overlap of single cell masks in consecutive images through deep learning video frame interpolation. As proof of principle, we used this approach to quantitatively image and compare cell growth and cell cycle regulation across the sexual life cycle of Saccharomyces cerevisiae. We developed a fluorescent reporter system based on a fluorescently labeled Whi5 protein, the yeast analog of mammalian Rb, and a new High-Cdk1 activity sensor, LiCHI, designed to report during DNA replication, mitosis, meiotic homologous recombination, meiosis I, and meiosis II. We found that cell growth preceded the exit from non-proliferative states such as mitotic G1, pre-meiotic G1, and the G0 spore state during germination. A decrease in the total cell concentration of Whi5 characterized the exit from non-proliferative states, which is consistent with a Whi5 dilution model. The nuclear accumulation of Whi5 was developmentally regulated, being at its highest during meiotic exit and spore formation. The temporal coordination of cell division and growth was not significantly different across three sexually reproducing generations. Our framework could be used to quantitatively characterize other single-cell eukaryotic life cycles that remain incompletely described. An off-the-shelf user interface Yeastvision provides free access to our image processing and single-cell tracking algorithms.

Lysosomal damage sensing and lysophagy initiation by SPG20-ITCH

Cells respond to lysosomal membrane permeabilization by membrane repair or selective macroautophagy of damaged lysosomes, termed lysophagy, but it is not fully understood how this decision is made. Here, we uncover a pathway in human cells that detects lipid bilayer perturbations in the limiting membrane of compromised lysosomes, which fail to be repaired, and then initiates ubiquitin-triggered lysophagy. We find that SPG20 binds the repair factor IST1 on damaged lysosomes and, importantly, integrates that with the detection of damage-associated lipid-packing defects of the lysosomal membrane. Detection occurs via sensory amphipathic helices in SPG20 before rupture of the membrane. If lipid-packing defects are extensive, such as during lipid peroxidation, SPG20 recruits and activates ITCH, which marks the damaged lysosome with lysine-63-linked ubiquitin chains to initiate lysophagy and thus triages the lysosome for destruction. With SPG20 being linked to neurodegeneration, these findings highlight the relevance of a coordinated lysosomal damage response for cellular homeostasis.

Tools and methods for high-throughput single-cell imaging with the mother machine

Despite much progress, image processing remains a significant bottleneck for high-throughput analysis of microscopy data. One popular platform for single-cell time-lapse imaging is the mother machine, which enables long-term tracking of microbial cells under precisely controlled growth conditions. While several mother machine image analysis pipelines have been developed in the past several years, adoption by a non-expert audience remains a challenge. To fill this gap, we implemented our own software, MM3, as a plugin for the multidimensional image viewer napari. napari-MM3 is a complete and modular image analysis pipeline for mother machine data, which takes advantage of the high-level interactivity of napari. Here, we give an overview of napari-MM3 and test it against several well-designed and widely used image analysis pipelines, including BACMMAN and DeLTA. Researchers often analyze mother machine data with custom scripts using varied image analysis methods, but a quantitative comparison of the output of different pipelines has been lacking. To this end, we show that key single-cell physiological parameter correlations and distributions are robust to the choice of analysis method. However, we also find that small changes in thresholding parameters can systematically alter parameters extracted from single-cell imaging experiments. Moreover, we explicitly show that in deep learning-based segmentation, 'what you put is what you get' (WYPIWYG) - that is, pixel-level variation in training data for cell segmentation can propagate to the model output and bias spatial and temporal measurements. Finally, while the primary purpose of this work is to introduce the image analysis software that we have developed over the last decade in our lab, we also provide information for those who want to implement mother machine-based high-throughput imaging and analysis methods in their research.

An experimental framework to assess biomolecular condensates in bacteria

High-resolution imaging of biomolecular condensates in living cells is essential for correlating their properties to those observed through in vitro assays. However, such experiments are limited in bacteria due to resolution limitations. Here we present an experimental framework that probes the formation, reversibility, and dynamics of condensate-forming proteins in Escherichia coli as a means to determine the nature of biomolecular condensates in bacteria. We demonstrate that condensates form after passing a threshold concentration, maintain a soluble fraction, dissolve upon shifts in temperature and concentration, and exhibit dynamics consistent with internal rearrangement and exchange between condensed and soluble fractions. We also discover that an established marker for insoluble protein aggregates, IbpA, has different colocalization patterns with bacterial condensates and aggregates, demonstrating its potential applicability as a reporter to differentiate the two in vivo. Overall, this framework provides a generalizable, accessible, and rigorous set of experiments to probe the nature of biomolecular condensates on the sub-micron scale in bacterial cells.

Genetic interaction mapping reveals functional relationships between peptidoglycan endopeptidases and carboxypeptidases

Peptidoglycan (PG) is the main component of the bacterial cell wall; it maintains cell shape while protecting the cell from internal osmotic pressure and external environmental challenges. PG synthesis is essential for bacterial growth and survival, and a series of PG modifications are required to allow expansion of the sacculus. Endopeptidases (EPs), for example, cleave the crosslinks between adjacent PG strands to allow the incorporation of newly synthesized PG. EPs are collectively essential for bacterial growth and must likely be carefully regulated to prevent sacculus degradation and cell death. However, EP regulation mechanisms are poorly understood. Here, we used TnSeq to uncover novel EP regulators in Vibrio cholerae. This screen revealed that the carboxypeptidase DacA1 (PBP5) alleviates EP toxicity. dacA1 is essential for viability on LB medium, and this essentiality was suppressed by EP overexpression, revealing that EP toxicity both mitigates, and is mitigated by, a defect in dacA1. A subsequent suppressor screen to restore viability of ΔdacA1 in LB medium identified hypomorphic mutants in the PG synthesis pathway, as well as mutations that promote EP activation. Our data thus reveal a more complex role of DacA1 in maintaining PG homeostasis than previously assumed.

Protocol to train a support vector machine for the automatic curation of bacterial cell detections in microscopy images

Manual curation of bacterial cell detections in microscopy images remains a time-consuming and laborious task. This work offers a comprehensive, step-by-step tutorial on training a support vector machine to autonomously distinguish between good and bad cell detections. Jupyter notebooks are included to perform feature extraction, labeling, and training of the machine learning model. This method can readily be incorporated into profiling pipelines aimed at extracting a multitude of features across large collections of individual cells, strains, and species. For complete details on the use and execution of this protocol, please refer to Govers et al.1.

A Foundation Model for Cell Segmentation

Cells are a fundamental unit of biological organization, and identifying them in imaging data - cell segmentation - is a critical task for various cellular imaging experiments. While deep learning methods have led to substantial progress on this problem, most models in use are specialist models that work well for specific domains. Methods that have learned the general notion of "what is a cell" and can identify them across different domains of cellular imaging data have proven elusive. In this work, we present CellSAM, a foundation model for cell segmentation that generalizes across diverse cellular imaging data. CellSAM builds on top of the Segment Anything Model (SAM) by developing a prompt engineering approach for mask generation. We train an object detector, CellFinder, to automatically detect cells and prompt SAM to generate segmentations. We show that this approach allows a single model to achieve human-level performance for segmenting images of mammalian cells (in tissues and cell culture), yeast, and bacteria collected across various imaging modalities. We show that CellSAM has strong zero-shot performance and can be improved with a few examples via few-shot learning. We also show that CellSAM can unify bioimaging analysis workflows such as spatial transcriptomics and cell tracking. A deployed version of CellSAM is available at https://cellsam.deepcell.org/.

Staphylococcus aureus AbcA transporter enhances persister formation under β-lactam exposure

We evaluated the role of Staphylococcus aureus AbcA transporter in bacterial persistence and survival following exposure to the bactericidal agents nafcillin and oxacillin at both the population and single-cell levels. We show that AbcA overexpression resulted in resistance to nafcillin but not oxacillin. Using distinct fluorescent reporters of cell viability and AbcA expression, we found that over 6-14 hours of persistence formation, the proportion of AbcA reporter-expressing cells assessed by confocal microscopy increased sixfold as cell viability reporters decreased. Similarly, single-cell analysis in a high-throughput microfluidic system found a strong correspondence between antibiotic exposure and AbcA reporter expression. Persister cells grown in the absence of antibiotics showed neither an increase in nafcillin MIC nor in abcA transcript levels, indicating that survival was not associated with stable mutational resistance or abcA overexpression. Furthermore, persister cell levels on exposure to 1×MIC and 25×MIC of nafcillin decreased in an abcA knockout mutant. Survivors of nafcillin and oxacillin treatment overexpressed transporter AbcA, contributing to an enrichment of the number of persisters during treatment with pump-substrate nafcillin but not with pump-non-substrate oxacillin, indicating that efflux pump expression can contribute selectively to the survival of a persister population.

PrestoCell: A persistence-based clustering approach for rapid and robust segmentation of cellular morphology in three-dimensional data

Light microscopy methods have continued to advance allowing for unprecedented analysis of various cell types in tissues including the brain. Although the functional state of some cell types such as microglia can be determined by morphometric analysis, techniques to perform robust, quick, and accurate measurements have not kept pace with the amount of imaging data that can now be generated. Most of these image segmentation tools are further burdened by an inability to assess structures in three-dimensions. Despite the rise of machine learning techniques, the nature of some biological structures prevents the training of several current day implementations. Here we present PrestoCell, a novel use of persistence-based clustering to segment cells in light microscopy images, as a customized Python-based tool that leverages the free multidimensional image viewer Napari. In evaluating and comparing PrestoCell to several existing tools, including 3DMorph, Omipose, and Imaris, we demonstrate that PrestoCell produces image segmentations that rival these solutions. In particular, our use of cell nuclei information resulted in the ability to correctly segment individual cells that were interacting with one another to increase accuracy. These benefits are in addition to the simplified graphically based user refinement of cell masks that does not require expensive commercial software licenses. We further demonstrate that PrestoCell can complete image segmentation in large samples from light sheet microscopy, allowing quantitative analysis of these large datasets. As an open-source program that leverages freely available visualization software, with minimum computer requirements, we believe that PrestoCell can significantly increase the ability of users without data or computer science expertise to perform complex image analysis.

Software Tools for 2D Cell Segmentation

Cell segmentation is an important task in the field of image processing, widely used in the life sciences and medical fields. Traditional methods are mainly based on pixel intensity and spatial relationships, but have limitations. In recent years, machine learning and deep learning methods have been widely used, providing more-accurate and efficient solutions for cell segmentation. The effort to develop efficient and accurate segmentation software tools has been one of the major focal points in the field of cell segmentation for years. However, each software tool has unique characteristics and adaptations, and no universal cell-segmentation software can achieve perfect results. In this review, we used three publicly available datasets containing multiple 2D cell-imaging modalities. Common segmentation metrics were used to evaluate the performance of eight segmentation tools to compare their generality and, thus, find the best-performing tool.