Label-free deep learning-based species classification of bacteria imaged by phase-contrast microscopy.

Reliable detection and classification of bacteria and other pathogens in the human body, animals, food, and water is crucial for improving and safeguarding public health. For instance, identifying the species and its antibiotic susceptibility is vital for effective bacterial infection treatment. Here we show that phase contrast time-lapse microscopy combined with deep learning is sufficient to classify four species of bacteria relevant to human health. The classification is performed on living bacteria and does not require fixation or staining, meaning that the bacterial species can be determined as the bacteria reproduce in a microfluidic device, enabling parallel determination of susceptibility to antibiotics. We assess the performance of convolutional neural networks and vision transformers, where the best model attained a class-average accuracy exceeding 98%. Our successful proof-of-principle results suggest that the methods should be challenged with data covering more species and clinically relevant isolates for future clinical use.

RegiSTORM: channel registration for multi-color stochastic optical reconstruction microscopy.

BACKGROUND: Stochastic optical reconstruction microscopy (STORM), a super-resolution microscopy technique based on single-molecule localizations, has become popular to characterize sub-diffraction limit targets. However, due to lengthy image acquisition, STORM recordings are prone to sample drift. Existing cross-correlation or fiducial marker-based algorithms allow correcting the drift within each channel, but misalignment between channels remains due to interchannel drift accumulating during sequential channel acquisition. This is a major drawback in multi-color STORM, a technique of utmost importance for the characterization of various biological interactions. RESULTS: We developed RegiSTORM, a software for reducing channel misalignment by accurately registering STORM channels utilizing fiducial markers in the sample. RegiSTORM identifies fiducials from the STORM localization data based on their non-blinking nature and uses them as landmarks for channel registration. We first demonstrated accurate registration on recordings of fiducials only, as evidenced by significantly reduced target registration error with all the tested channel combinations. Next, we validated the performance in a more practically relevant setup on cells multi-stained for tubulin. Finally, we showed that RegiSTORM successfully registers two-color STORM recordings of cargo-loaded lipid nanoparticles without fiducials, demonstrating the broader applicability of this software. CONCLUSIONS: The developed RegiSTORM software was demonstrated to be able to accurately register multiple STORM channels and is freely available as open-source (MIT license) at https://github.com/oystein676/RegiSTORM.git and https://doi.org/10.5281/zenodo.5509861 (archived), and runs as a standalone executable (Windows) or via Python (Mac OS, Linux).

Occurrence of late-apoptotic symptoms in porcine preimplantation embryos upon exposure of oocytes to perfluoroalkyl substances (PFASs) under in vitro meiotic maturation.

The objectives of this study were to evaluate the effect of perfluoroalkyl substances on early embryonic development and apoptosis in blastocysts using a porcine in vitro model. Porcine oocytes (N = 855) collected from abattoir ovaries were subjected to perfluorooctane sulfonic acid (PFOS) (0.1 µg/ml) and perfluorohexane sulfonic acid (PFHxS) (40 µg/ml) during in vitro maturation (IVM) for 45 h. The gametes were then fertilized and cultured in vitro, and developmental parameters were recorded. After 6 days of culture, resulting blastocysts (N = 146) were stained using a terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay and imaged as stacks using confocal laser scanning microscopy. Proportion of apoptotic cells as well as total numbers of nuclei in each blastocyst were analyzed using objective image analysis. The experiment was run in 9 replicates, always with a control present. Effects on developmental parameters were analyzed using logistic regression, and effects on apoptosis and total numbers of nuclei were analyzed using linear regression. Higher cell count was associated with lower proportion of apoptotic cells, i.e., larger blastocysts contained less apoptotic cells. Upon PFAS exposure during IVM, PFHxS tended to result in higher blastocyst rates on day 5 post fertilization (p = 0.07) and on day 6 post fertilization (p = 0.05) as well as in higher apoptosis rates in blastocysts (p = 0.06). PFHxS resulted in higher total cell counts in blastocysts (p = 0.002). No effects attributable to the concentration of PFOS used here was seen. These findings add to the evidence that some perfluoroalkyl substances may affect female reproduction. More studies are needed to better understand potential implications for continued development as well as for human health.

mafba and mafbb differentially regulate lymphatic endothelial cell migration in topographically distinct manners.

Lymphangiogenesis, formation of lymphatic vessels from pre-existing vessels, is a dynamic process that requires cell migration. Regardless of location, migrating lymphatic endothelial cell (LEC) progenitors probe their surroundings to form the lymphatic network. Lymphatic-development regulation requires the transcription factor MAFB in different species. Zebrafish Mafba, expressed in LEC progenitors, is essential for their migration in the trunk. However, the transcriptional mechanism that orchestrates LEC migration in different lymphatic endothelial beds remains elusive. Here, we uncover topographically different requirements of the two paralogs, Mafba and Mafbb, for LEC migration. Both mafba and mafbb are necessary for facial lymphatic development, but mafbb is dispensable for trunk lymphatic development. On the molecular level, we demonstrate a regulatory network where Vegfc-Vegfd-SoxF-Mafba-Mafbb is essential in facial lymphangiogenesis. We identify that mafba and mafbb tune the directionality of LEC migration and vessel morphogenesis that is ultimately necessary for lymphatic function.

Bovine oocyte exposure to perfluorohexane sulfonate (PFHxS) induces phenotypic, transcriptomic, and DNA methylation changes in resulting embryos in vitro.

Knowledge on the effects of perfluorohexane sulfonate (PFHxS) on ovarian function is limited. In the current study, we investigated the sensitivity of oocytes to PFHxS during in vitro maturation (IVM), including consequences on embryo development at the morphological, transcriptomic, and epigenomic levels. Bovine cumulus-oocyte complexes (COCs) were exposed to PFHxS during 22 h IVM. Following fertilisation, developmental competence was recorded until day 8 of culture. Two experiments were conducted: 1) exposure of COCs to 0.01 µg mL-1 - 100 µg mL-1 PFHxS followed by confocal imaging to detect neutral lipids and nuclei, and 2) exposure of COCs to 0.1 µg mL-1 PFHxS followed by analysis of transcriptomic and DNA methylation changes in blastocysts. Decreased oocyte developmental competence was observed upon exposure to ≥ 40 µg mL-1 PFHxS and altered lipid distribution was observed in the blastocysts upon exposure to 1-10 µg mL-1 PFHxS (not observed at lower or higher concentrations). Transcriptomic data showed that genes affected by 0.1 µg mL-1 PFHxS were enriched for pathways related to increased synthesis and production of reactive oxygen species. Enrichment for peroxisome proliferator-activated receptor-γ and oestrogen pathways was also observed. Genes linked to DNA methylation changes were enriched for similar pathways. In conclusion, exposure of the bovine oocyte to PFHxS during the narrow window of IVM affected subsequent embryonic development, as reflected by morphological and molecular changes. This suggests that PFHxS interferes with the final nuclear and cytoplasmic maturation of the oocyte leading to decreased developmental competence to blastocyst stage.

Increased apoptosis, reduced Wnt/ß-catenin signaling, and altered tail development in zebrafish embryos exposed to a human-relevant chemical mixture.

A wide variety of anthropogenic chemicals is detected in humans and wildlife and the health effects of various chemical exposures are not well understood. Early life stages are generally the most susceptible to chemical disruption and developmental exposure can cause disease in adulthood, but the mechanistic understanding of such effects is poor. Within the EU project EDC-MixRisk, a chemical mixture (Mixture G) was identified in the Swedish pregnancy cohort SELMA by the inverse association between levels in women at around gestational week ten with birth weight of their children. This mixture was composed of mono-ethyl phthalate, mono-butyl phthalate, mono-benzyl phthalate, mono-ethylhexyl phthalate, mono-isononyl phthalate, triclosan, perfluorohexane sulfonate, perfluorooctanoic acid, and perfluorooctane sulfonate. In a series of experimental studies, we characterized effects of Mixture G on early development in zebrafish models. Here, we studied apoptosis and Wnt/ß-catenin signaling which are two evolutionarily conserved signaling pathways of crucial importance during development. We determined effects on apoptosis by measuring TUNEL staining, caspase-3 activity, and acridine orange staining in wildtype zebrafish embryos, while Wnt/ß-catenin signaling was assayed using a transgenic line expressing an EGFP reporter at ß-catenin-regulated promoters. We found that Mixture G increased apoptosis, suppressed Wnt/ß-catenin signaling in the caudal fin, and altered the shape of the caudal fin at water concentrations only 20-100 times higher than the geometric mean serum concentration in the human cohort. These findings call for awareness that pollutant mixtures like mixture G may interfere with a variety of developmental processes, possibly resulting in adverse health effects.

SynQuant: an automatic tool to quantify synapses from microscopy images.

MOTIVATION: Synapses are essential to neural signal transmission. Therefore, quantification of synapses and related neurites from images is vital to gain insights into the underlying pathways of brain functionality and diseases. Despite the wide availability of synaptic punctum imaging data, several issues are impeding satisfactory quantification of these structures by current tools. First, the antibodies used for labeling synapses are not perfectly specific to synapses. These antibodies may exist in neurites or other cell compartments. Second, the brightness of different neurites and synaptic puncta is heterogeneous due to the variation of antibody concentration and synapse-intrinsic differences. Third, images often have low signal to noise ratio due to constraints of experiment facilities and availability of sensitive antibodies. These issues make the detection of synapses challenging and necessitates developing a new tool to easily and accurately quantify synapses. RESULTS: We present an automatic probability-principled synapse detection algorithm and integrate it into our synapse quantification tool SynQuant. Derived from the theory of order statistics, our method controls the false discovery rate and improves the power of detecting synapses. SynQuant is unsupervised, works for both 2D and 3D data, and can handle multiple staining channels. Through extensive experiments on one synthetic and three real datasets with ground truth annotation or manually labeling, SynQuant was demonstrated to outperform peer specialized unsupervised synapse detection tools as well as generic spot detection methods. AVAILABILITY AND IMPLEMENTATION: Java source code, Fiji plug-in, and test data are available at https://github.com/yu-lab-vt/SynQuant. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

In situ quantification of individual mRNA transcripts in melanocytes discloses gene regulation of relevance to speciation.

Functional validation of candidate genes involved in adaptation and speciation remains challenging. Here, we exemplify the utility of a method quantifying individual mRNA transcripts in revealing the molecular basis of divergence in feather pigment synthesis during early-stage speciation in crows. Using a padlock probe assay combined with rolling circle amplification, we quantified cell-type-specific gene expression in the histological context of growing feather follicles. Expression of Tyrosinase Related Protein 1 (TYRP1), Solute Carrier Family 45 member 2 (SLC45A2) and Hematopoietic Prostaglandin D Synthase (HPGDS) was melanocyte-limited and significantly reduced in follicles from hooded crow, explaining the substantially lower eumelanin content in grey versus black feathers. The central upstream Melanocyte Inducing Transcription Factor (MITF) only showed differential expression specific to melanocytes - a feature not captured by bulk RNA-seq. Overall, this study provides insight into the molecular basis of an evolutionary young transition in pigment synthesis, and demonstrates the power of histologically explicit, statistically substantiated single-cell gene expression quantification for functional genetic inference in natural populations.

Human Immunodeficiency Virus-Infected Women Have High Numbers of CD103-CD8+ T Cells Residing Close to the Basal Membrane of the Ectocervical Epithelium.

Background: Genital mucosa is the main portal of entry for various incoming pathogens, including human immunodeficiency virus (HIV), hence it is an important site for host immune defenses. Tissue-resident memory T (TRM) cells defend tissue barriers against infections and are characterized by expression of CD103 and CD69. In this study, we describe the composition of CD8+ TRM cells in the ectocervix of healthy and HIV-infected women. Methods: Study samples were collected from healthy Swedish and Kenyan HIV-infected and uninfected women. Customized computerized image-based in situ analysis was developed to assess the ectocervical biopsies. Genital mucosa and blood samples were assessed by flow cytometry. Results: Although the ectocervical epithelium of healthy women was populated with bona fide CD8+ TRM cells (CD103+CD69+), women infected with HIV displayed a high frequency of CD103-CD8+ cells residing close to their epithelial basal membrane. Accumulation of CD103-CD8+ cells was associated with chemokine expression in the ectocervix and HIV viral load. CD103+CD8+ and CD103-CD8+ T cells expressed cytotoxic effector molecules in the ectocervical epithelium of healthy and HIV-infected women. In addition, women infected with HIV had decreased frequencies of circulating CD103+CD8+ T cells. Conclusions: Our data provide insight into the distribution of CD8+ TRM cells in human genital mucosa, a critically important location for immune defense against pathogens, including HIV.

Quantitative high-content/high-throughput microscopy analysis of lipid droplets in subject-specific adipogenesis models.

Neutral lipids packed in lipid droplets (LDs) are essential as a source of fuel for organisms, and specialized storing cells, the adipocytes, provide a buffer for energy variations. Many modern-society-disorders are connected with excess accumulation or deficiency of LDs in adipose tissue. Intracellular LD number and size distribution reflect the tissue conditions, while the associated mechanisms and genes rs are still poorly understood. Large-scale genetic screens using human in vitro differentiated primary adipocytes require cell samples donated from many patients. The heterogeneity appearing between donors highlighted the need for high-throughput methods robust to individual variations. Previous image analysis algorithms failed to handle individual LDs, but focused on averages, hiding population heterogeneity. We present a new high-content analysis (HCA) technique for analysis of fat cell metabolism using data from a large-scale RNAi screen including images of more than 500 k in vitro differentiated adipocytes from three donors. The RNAi-based suppression of Perilipin 1 (PLIN1), a protein involved in the adipocyte lipid metabolism, served as a positive control, while cells treated with randomized RNA served as negative controls. We validate our segmentation by comparing our results to those of previously published methods: We also evaluate the discriminative power of different morphological features describing LD size distribution. Classification of cells as containing few large or many small LDs followed by calculating the percentage of cells in each class proved to discriminate the positive PLIN1-suppressed phenotype from the untreated negative control with an area under the receiver operating characteristic curve of 0.98. The results suggest that this HCA method offers improved segmentation and classification accuracy, and can, thus, be utilized to quantify changes in LD metabolism in response to treatment in many cell models relevant to a variety of diseases. © 2017 International Society for Advancement of Cytometry.

Automated Training of Deep Convolutional Neural Networks for Cell Segmentation.

Deep Convolutional Neural Networks (DCNN) have recently emerged as superior for many image segmentation tasks. The DCNN performance is however heavily dependent on the availability of large amounts of problem-specific training samples. Here we show that DCNNs trained on ground truth created automatically using fluorescently labeled cells, perform similar to manual annotations.

Global gray-level thresholding based on object size.

In this article, we propose a fast and robust global gray-level thresholding method based on object size, where the selection of threshold level is based on recall and maximum precision with regard to objects within a given size interval. The method relies on the component tree representation, which can be computed in quasi-linear time. Feature-based segmentation is especially suitable for biomedical microscopy applications where objects often vary in number, but have limited variation in size. We show that for real images of cell nuclei and synthetic data sets mimicking fluorescent spots the proposed method is more robust than all standard global thresholding methods available for microscopy applications in ImageJ and CellProfiler. The proposed method, provided as ImageJ and CellProfiler plugins, is simple to use and the only required input is an interval of the expected object sizes. © 2016 International Society for Advancement of Cytometry.

Compaction of rolling circle amplification products increases signal integrity and signal-to-noise ratio.

Rolling circle amplification (RCA) for generation of distinct fluorescent signals in situ relies upon the self-collapsing properties of single-stranded DNA in commonly used RCA-based methods. By introducing a cross-hybridizing DNA oligonucleotide during rolling circle amplification, we demonstrate that the fluorophore-labeled RCA products (RCPs) become smaller. The reduced size of RCPs increases the local concentration of fluorophores and as a result, the signal intensity increases together with the signal-to-noise ratio. Furthermore, we have found that RCPs sometimes tend to disintegrate and may be recorded as several RCPs, a trait that is prevented with our cross-hybridizing DNA oligonucleotide. These effects generated by compaction of RCPs improve accuracy of visual as well as automated in situ analysis for RCA based methods, such as proximity ligation assays (PLA) and padlock probes.

Automatic segmentation of intra-abdominal and subcutaneous adipose tissue in 3D whole mouse MRI.

PURPOSE: To fully automate intra-abdominal (IAT) and total adipose tissue (TAT) segmentation in mice to replace tedious and subjective manual segmentation. MATERIALS AND METHODS: A novel transform codes each voxel with the radius of the narrowest passage on the widest possible three-dimensional (3D) path to any voxel in the target object to select appropriate IAT seed points. Then competitive region growing is performed on a distance transform of the fat mask such that competing classes meet at narrow passages effectively segmenting the IAT and subcutaneous adipose compartments. Fully automatic segmentations were conducted on 32 3D mouse images independent to those used for algorithm development. RESULTS: Automatic processing worked on all 32 images and took 28 s on a 3.6 GHz Pentium computer with 2.0 GB RAM. Manual segmentation by an experienced operator typically took 1 h per 3D image. The correlation coefficients between manual and automated segmentation of TAT and IAT were 0.97 and 0.94, respectively. CONCLUSION: The fully automatic method correlates well with manual segmentation and dramatically speeds up segmentation allowing MRI to be used in the anti-obesity drug discovery pipeline.