3D Microscopy Images Research Articles

FC 038CRESCENTS DERIVE FROM SINGLE PODOCYTE PROGENITORS AND A DRUG ENHANCING THEIR DIFFERENTIATION ATTENUATES RAPIDLY PROGRESSIVE GLOMERULONEPHRITIS

Abstract Background and Aims Rapidly progressive glomerulonephritis (RPGN) encompasses a group of diverse disorders characterized by the presence of massive hyperplasia of parietal epithelial cells (PEC) as the main histopathological lesion at kidney biopsy. It is associated with a rapid decline in kidney function referred to altogether as rapidly progressive glomerulonephritis. Typically, crescent formation is the consequence of diverse upstream pathomechanisms involving the specific activation of PEC. PEC normally reside peacefully along Bowman capsule and represent in part renal progenitor cells (RPC). Previous studies observed RPC markers in crescents from patients with different types of glomerulonephritis. Similarities between stem cell niches of bone marrow and kidney, prompted us to hypothesized that crescents result from monoclonal expansion of a single RPC clone conceptually similar to monoclonal diseases originating from hematopoietic stem cells. According to this analogy, we further hypothesized that drugs known to cure monoclonal disease of the hematopoietic stem cells by enforcing their terminal differentiation could also attenuate crescentic glomerulonephritis. Method To address this hypothesis, we established a RPGN disease model in a conditional transgenic mouse based on the mT/mG and the Confetti reporter that allows lineage tracing and clonal analysis of RPCs. Animals were treated with known pharmacological inhibitors of clonal stem cell proliferation in myeloproliferative disorders. Crescentic lesions were characterized by super-resolution STED microscopy. Finally, we employed single cell RNA sequencing of human renal progenitor cultures to identify the immature progenitor subset-generating crescent in human to identify putative new biomarker(s) of RPNG to validate in biopsy of patients. Results We observed that the crescentic lesions originated from the clonal expansion of single RPC, thus suggesting a clonal stem cell disorder. Therefore, we administrated a series of drugs known to ameliorates myeloproliferative neoplasms to our RPGN mouse model as potential therapeutic agents. In particular, treatment with one of the compounds induced a reduction in both proteinuria and crescent formation. 3D confocal microscopy and STED super-resolution imaging of glomeruli showed that this compound turned the uncontrolled hyperplasia of a specific immature PEC subset into a controlled differentiation into new podocytes thereby restoring the injured glomerular filtration barrier. Single cell RNA sequencing of human renal progenitor cultures identified a new marker of the crescent-generating progenitor cells. Expression of this marker in biopsies of patients with rapidly progressive glomerulonephritis associated with progression toward end stage kidney disease. Treatment of human PEC with the drug that in in vivo experiments showed a therapeutic effect on RPGN reduced proliferation of the immature progenitor subset promoting their differentiation into podocytes. Conclusion These results demonstrate that glomerular hyperplastic lesions derive from clonal amplification of a RPC subset and that shifting proliferation to podocyte differentiation reverses crescent formation and improves clinical outcome.

The nanoscale organization of the Wnt signaling integrator Dishevelled in the vegetal cortex domain of an egg and early embryo.

Canonical Wnt/β-catenin (cWnt) signaling is a crucial regulator of development and Dishevelled (Dsh/Dvl) functions as an integral part of this pathway by linking Wnt binding to the Frizzled:LRP5/6 receptor complex with β-catenin-stimulated gene expression. In many cell types Dsh has been localized to ill-defined cytoplasmic puncta, however in sea urchin eggs and embryos confocal fluorescence microscopy has shown that Dsh is localized to puncta present in a novel and development-essential vegetal cortex domain (VCD). In the present study, we used super-resolution light microscopy and platinum replica transmission electron microscopy (TEM) to provide the first views of the ultrastructural organization of Dsh within the sea urchin VCD. 3D structured illumination microscopy (SIM) imaging of isolated egg cortices demonstrated the graded distribution of Dsh in the VCD, whereas higher resolution stimulated emission depletion (STED) imaging revealed that some individual Dsh puncta consisted of more than one fluorescent source. Platinum replica immuno-TEM localization showed that Dsh puncta on the cytoplasmic face of the plasma membrane consisted of aggregates of pedestal-like structures each individually labeled with the C-terminus specific Dsh antibody. These aggregates were resistant to detergent extraction and treatment with drugs that disrupt actin filaments or inhibit myosin II contraction, and coexisted with the first cleavage actomyosin contractile ring. These results confirm and extend previous studies and reveal, for the first time in any cell type, the nanoscale organization of plasma membrane tethered Dsh. Our current working hypothesis is that these Dsh pedestals represent a prepositioned scaffold organization that is important for the localized activation of the cWnt pathway at the sea urchin vegetal pole. These observations in sea urchins may also be relevant to the submembranous Dsh puncta present in other eggs and embryos.

Segmentor: a tool for manual refinement of 3D microscopy annotations

BackgroundRecent advances in tissue clearing techniques, combined with high-speed image acquisition through light sheet microscopy, enable rapid three-dimensional (3D) imaging of biological specimens, such as whole mouse brains, in a matter of hours. Quantitative analysis of such 3D images can help us understand how changes in brain structure lead to differences in behavior or cognition, but distinguishing densely packed features of interest, such as nuclei, from background can be challenging. Recent deep learning-based nuclear segmentation algorithms show great promise for automated segmentation, but require large numbers of accurate manually labeled nuclei as training data.ResultsWe present Segmentor, an open-source tool for reliable, efficient, and user-friendly manual annotation and refinement of objects (e.g., nuclei) within 3D light sheet microscopy images. Segmentor employs a hybrid 2D-3D approach for visualizing and segmenting objects and contains features for automatic region splitting, designed specifically for streamlining the process of 3D segmentation of nuclei. We show that editing simultaneously in 2D and 3D using Segmentor significantly decreases time spent on manual annotations without affecting accuracy as compared to editing the same set of images with only 2D capabilities.ConclusionsSegmentor is a tool for increased efficiency of manual annotation and refinement of 3D objects that can be used to train deep learning segmentation algorithms, and is available at https://www.nucleininja.org/ and https://github.com/RENCI/Segmentor.

Relational Classification of Biological Cells in Microscopy Images

We investigate the relational classification of biological cells in 2D microscopy images. Rather than treating each cell image independently, we investigate whether and how the neighborhood information of a cell can be informative for its prediction. We propose a Relational Long Short-Term Memory (R-LSTM) algorithm, coupled with auto-encoders and convolutional neural networks, that can learn from both annotated and unlabeled microscopy images and that can utilize both the local and neighborhood information to perform an improved classification of biological cells. Experimental results on both synthetic and real datasets show that R-LSTM performs comparable to or better than six baselines.

Nested U-Net Architecture Based Image Segmentation for 3D Neuron Reconstruction

Digital reconstruction of neurons is a critical step in studying neuronal morphology and exploring the working mechanism of the brain. In recent years, the focus of neuronal morphology reconstruction has gradually shifted from single neurons to multiple neurons in a whole brain. Microscopic images of a whole brain often have low signal-to-noise-ratio, discontinuous neuron fragments or weak neuron signals. It is very difficult to segment neuronal signals from the background of these images, which is the first step of most automatic reconstruction algorithms. In this study, we propose a Nested U-Net based Ultra-Tracer model (NUNU-Tracer) for better multiple neurons image segmentation and morphology reconstruction. The NUNU-Tracer utilizes nested U-Net (UNet++) deep network to segment 3D neuron images, reconstructs neuron morphologies under the framework of the Ultra-Tracer and prunes branches of noncurrent tracing neurons. The 3D UNet++ takes a 3D microscopic image as its input, and uses scale-space distance transform and linear fusion strategy to generate the segmentation maps for voxels in the image. It is capable of removing noise, repairing broken neurite patterns and enhancing neuronal signals. We evaluate the performance of the 3D UNet++ for image segmentation and NUNU-Tracer for neuron morphology reconstruction on image blocks and neurons, respectively. Experimental results show that they significantly improve the accuracy and length of neuron reconstructions.

FRC-QE: a robust and comparable 3D microscopy image quality metric for cleared organoids.

SummaryHere, we propose Fourier ring correlation-based quality estimation (FRC-QE) as a new metric for automated image quality estimation in 3D fluorescence microscopy acquisitions of cleared organoids that yields comparable measurements across experimental replicates, clearing protocols and works for different microscopy modalities.Availability and implementationFRC-QE is written in ImgLib2/Java and provided as an easy-to-use and macro-scriptable plugin for Fiji. Code, documentation, sample images and further information can be found under https://github.com/PreibischLab/FRC-QE.Supplementary informationSupplementary data are available at Bioinformatics online.

DeepS: a web server for image optical sectioning and super resolution microscopy based on a deep learning framework.

Microscopy technology plays important roles in many biological research fields. Solvent-cleared brain high-resolution (HR) 3D image reconstruction is an important microscopy application. However, 3D microscopy image generation is time-consuming and expensive. Therefore, we have developed a deep learning framework (DeepS) for both image optical sectioning and super resolution microscopy. Using DeepS to perform super resolution solvent-cleared mouse brain microscopy 3D image yields improved performance in comparison with the standard image processing workflow. We have also developed a web server to allow online usage of DeepS. Users can train their own models with only one pair of training images using the transfer learning function of the web server. http://deeps.cibr.ac.cn. Supplementary data are available at Bioinformatics online.

Fractography of clinical failures of indirect resin composite endocrown and overlay restorations

ObjectivesCompare failure modes and fracture origins using fractography on recovered clinically fractured parts of indirect resin composite endocrowns and overlay restorations on endodontically treated teeth (ETT). MethodsFour endocrowns (3 molars, 1 premolar) and one overlay (molar) adhesively luted on ETT were recovered after fracturing during function. The time in service ranged between 4 and 48 months. The composite materials were (i) CAD/CAM LAVA Ultimate (N = 1), (ii) Premise Indirect (N = 2), and (iii) Colombus (N = 2). Fractography was performed by means of digital microscopy and SEM. Occlusal surfaces were checked for signs of fatigue degradation and contact wear. Cuspal plane angles were measured from profiles obtained from 3D digital microscope images with respect to the horizontal plane of the occlusal central crown groove. ResultsAll five cases showed a wedge-opening mode I fracture, splitting the crown and tooth in two parts through the crown’s central groove. Classic brittle fracture features (arrest lines, twist and wake hackle) were easily identified on the fracture surfaces. Multiple origins were located along the central groove in conjunction with the presence of fatigue cracks. Contact wear surfaces showed pitting and cracking. Cuspal plane angles were around 30–35°, except a 50° palatal cusp slope for the Lava Ultimate overlay. SignificanceFractography on clinical fractures of resin composites was enlightening. Occlusal surface fatigue degradation from cyclic loading, mode I fracture from applied mastication forces on cuspal planes, and stress concentration within the crown’s central groove, indicate limitations of use of these materials for endocrowns in posterior teeth.

PartSeg: a tool for quantitative feature extraction from 3D microscopy images for dummies

BackgroundBioimaging techniques offer a robust tool for studying molecular pathways and morphological phenotypes of cell populations subjected to various conditions. As modern high-resolution 3D microscopy provides access to an ever-increasing amount of high-quality images, there arises a need for their analysis in an automated, unbiased, and simple way. Segmentation of structures within the cell nucleus, which is the focus of this paper, presents a new layer of complexity in the form of dense packing and significant signal overlap. At the same time, the available segmentation tools provide a steep learning curve for new users with a limited technical background. This is especially apparent in the bulk processing of image sets, which requires the use of some form of programming notation.ResultsIn this paper, we present PartSeg, a tool for segmentation and reconstruction of 3D microscopy images, optimised for the study of the cell nucleus. PartSeg integrates refined versions of several state-of-the-art algorithms, including a new multi-scale approach for segmentation and quantitative analysis of 3D microscopy images. The features and user-friendly interface of PartSeg were carefully planned with biologists in mind, based on analysis of multiple use cases and difficulties encountered with other tools, to offer an ergonomic interface with a minimal entry barrier. Bulk processing in an ad-hoc manner is possible without the need for programmer support. As the size of datasets of interest grows, such bulk processing solutions become essential for proper statistical analysis of results. Advanced users can use PartSeg components as a library within Python data processing and visualisation pipelines, for example within Jupyter notebooks. The tool is extensible so that new functionality and algorithms can be added by the use of plugins. For biologists, the utility of PartSeg is presented in several scenarios, showing the quantitative analysis of nuclear structures.ConclusionsIn this paper, we have presented PartSeg which is a tool for precise and verifiable segmentation and reconstruction of 3D microscopy images. PartSeg is optimised for cell nucleus analysis and offers multi-scale segmentation algorithms best-suited for this task. PartSeg can also be used for the bulk processing of multiple images and its components can be reused in other systems or computational experiments.

3D Neuron Microscopy Image Segmentation via the Ray-Shooting Model and a DC-BLSTM Network.

The morphology reconstruction (tracing) of neurons in 3D microscopy images is important to neuroscience research. However, this task remains very challenging because of the low signal-to-noise ratio (SNR) and the discontinued segments of neurite patterns in the images. In this paper, we present a neuronal structure segmentation method based on the ray-shooting model and the Long Short-Term Memory (LSTM)-based network to enhance the weak-signal neuronal structures and remove background noise in 3D neuron microscopy images. Specifically, the ray-shooting model is used to extract the intensity distribution features within a local region of the image. And we design a neural network based on the dual channel bidirectional LSTM (DC-BLSTM) to detect the foreground voxels according to the voxel-intensity features and boundary-response features extracted by multiple ray-shooting models that are generated in the whole image. This way, we transform the 3D image segmentation task into multiple 1D ray/sequence segmentation tasks, which makes it much easier to label the training samples than many existing Convolutional Neural Network (CNN) based 3D neuron image segmentation methods. In the experiments, we evaluate the performance of our method on the challenging 3D neuron images from two datasets, the BigNeuron dataset and the Whole Mouse Brain Sub-image (WMBS) dataset. Compared with the neuron tracing results on the segmented images produced by other state-of-the-art neuron segmentation methods, our method improves the distance scores by about 32% and 27% in the BigNeuron dataset, and about 38% and 27% in the WMBS dataset.

YI 1.7 Transmural Quantification of Murine Vascular Smooth Muscle Cell Density Distribution from 3D Microscopy Images

PurposeInvestigating the biomechanical role of smooth muscle cells (SMCs) in arteries requires knowledge of their structural distributions. Compared to histology, 3D microscopy offers non-destructive ex vivo imaging under realistic conditions [1]. Robust 3D segmentation of SMCs, however, is challenging. We propose a method for automatic SMC quantification, and assessed its potential using a murine SMC apoptosis model.MethodsAfter euthanasia, carotid arteries (control and with induced SMC apoptosis: SM22α-hDTR [2]) were excised and mounted between micropipettes (Figure A). Nuclei were stained with SYTO41. Arteries were imaged using two-photon microscopy [1], while stretched to in vivo length and pressurised to 100 mmHg (Figure B). Image stacks were processed as follows: 1) deconvolution; 2) nuclei segmentation using vesselness filtering [3,4] (Figure C); 3) cylindrical coordinate system identification; 4) splitting of coincident nuclei, based on cores defined from groups of neighbouring voxels with similar orientations [3] (Figure D and E); 5) cylindrical coordinate system re-identification; and 6) cell density-distribution quantification (Figure F). Segmentation performance was assessed by comparing with manual cell counts.ResultsFigure E demonstrates the method’s ability to split undersegmented coinciding nuclei. Cell counts were lower in SM22α-hDTR than in control; algorithm-derived counts were comparable to manual (Figure F). The control sample showed multiple SMC layers, while the SM22α-hDTR sample showed a single SMC layer (Figure F), which was confirmed visually.ConclusionWe developed a precise tool to quantify SMC distributions in ex vivo murine arteries, to facilitate quantitative modelling of SMC biomechanics. We intend to expand the current approach to address cell orientation, shape, and size.Figure(A) Imaging set-up illustrating acquisition of z-stack of slices. (B) Example slice of 3D stack; cell nuclei are shown in blue while elastin fibres are shown in red. (C) Segmentation results from vesselness filtering of example image stack, colours indicate separated nuclei (step 2, Methods). (D) Coinciding nuclei, corresponding with the left orange nuclei in C, shown in grey, with cell cores shown in red and blue (step 4, Methods). (E) Coinciding nuclei splitting results of nuclei shown in C. Nonsplit nuclei are shown in grey, while split nuclei are shown in red and blue. (F) Transmural SMC densities and cell counts for one control and one SMC apoptosis sample; manual cell counts are given between parentheses.

3D convolutional neural networks-based segmentation to acquire quantitative criteria of the nucleus during mouse embryogenesis

During embryogenesis, cells repeatedly divide and dynamically change their positions in three-dimensional (3D) space. A robust and accurate algorithm to acquire the 3D positions of the cells would help to reveal the mechanisms of embryogenesis. To acquire quantitative criteria of embryogenesis from time-series 3D microscopic images, image processing algorithms such as segmentation have been applied. Because the cells in embryos are considerably crowded, an algorithm to segment individual cells in detail and accurately is needed. To quantify the nuclear region of every cell from a time-series 3D fluorescence microscopic image of living cells, we developed QCANet, a convolutional neural network-based segmentation algorithm for 3D fluorescence bioimages. We demonstrated that QCANet outperformed 3D Mask R-CNN, which is currently considered as the best algorithm of instance segmentation. We showed that QCANet can be applied not only to developing mouse embryos but also to developing embryos of two other model species. Using QCANet, we were able to extract several quantitative criteria of embryogenesis from 11 early mouse embryos. We showed that the extracted criteria could be used to evaluate the differences between individual embryos. This study contributes to the development of fundamental approaches for assessing embryogenesis on the basis of extracted quantitative criteria.

Cover Feature: Determination of Rate Constants and Reaction Orders of Vanadium‐Ion Kinetics on Carbon Fiber Electrodes. (ChemElectroChem 21/2020)

The Cover Feature shows a 3D microscopic image of a single carbon fiber microelectrode. The fibers are obtained from carbon felt materials, which are commonly used as electrodes in vanadium redox flow batteries. As the porous structure makes detailed investigations of the vanadium ion kinetics difficult, studies on single fiber electrodes can contribute to a better understanding of the electrode processes. More information can be found in the Article by I. Kroner et al.

Automatic evaluation system for bulk defects in optics

In the high frequency laser system, the internal defects of optics will affect their damage threshold and downstream modulation, and even lead to subsequent surface damage. In order to efficiently and reliably evaluate the internal defects in optics, an automatic evaluation method of scattering defects in optics is proposed, and the corresponding three-dimensional (3D) precision measurement system is established. The detection principle, image acquisition, image processing and bulk defects (BDs) information extraction methods of the system are studied. Based on a 3D dark-field microscopic imaging (DFMI) technology, samples are scanned with multi-dimensional motion devices. Meanwhile, the scattered beam from the BD was collected by a detector, and finally combined with the digital image processing technology, the acquired images are processed in real time. Through real-time focusing and improved 3D reconstruction algorithm, the images with full aperture BD features are obtained, and the position and size information of defects are extracted. The system has detected the BD inside the fused silica samples. The results show that our system can distinguish the 612.7 lp/mm. Furthermore, the 3D distributions of the BD were also reconstructed.

Enhanced Antibiofilm Effects of N2 Plasma-Treated Buffer Combined with Antimicrobial Hexapeptides Against Plant Pathogens.

Suppression of pathogenic bacterial growth to increase food and agricultural productivity is important. We previously developed novel hexapeptides (KCM12 and KCM21) with antimicrobial activities against various phytopathogenic bacteria and N2 plasma-treated buffer (NPB) as an alternative method for bacterial inactivation and as an antibiofilm agent of crops. Here, we developed an enhanced antibiofilm method based on antimicrobial hexapeptides with N2 plasma-treated buffer against plant pathogens. Our results demonstrated that hexapeptides effectively inhibited the growth of Pseudomonas syringae pv. tomato DC3000 (Pst DC3000) and the biofilm it formed. Potent biofilm formation-inhibiting effects of hexapeptides were observed at concentrations of above 20 µM, and samples treated with hexapeptide above 100 µM reduced the ability of the bacteria to produce biofilm by 80%. 3D confocal laser scanning microscopy imaging data revealed that the antimicrobial activity of hexapeptides was enough to affect the cells embedded inside the biofilm. Finally, combination treatment with NPB and antimicrobial hexapeptides increased the antibiofilm effect compared with the effect of single processing against multilayered plant pathogen biofilms. These findings show that the combination of hexapeptides and NPB can be potentially applied for improving crop production.

Tribological Behaviour of Additively Manufactured Fiber-Reinforced Thermoplastic Composites in Various Environments.

Polymer composites with increased utility properties are becoming competition for conventional materials, in conjunction with additive manufacturing techniques. The aim of this study was to evaluate tribological characteristics of fibrous composites produced in fused deposition modeling (FDM) with the use of an innovative head with symmetrical feeding of the matrix material. Analysis of the influence of composite-forming temperature on their tribological properties allowed the determining of the optimal printing process parameters for this group of composites. Significant differences in the friction process of the same reinforced materials were observed in dry and wet environments. Fibrous composites showed 10 times lower wear intensity as well as two times lower friction value in water than in air. Research shows friction in the water environment ensures more even wear of the surface of the composites involved in the work. The article contains 3D microscopic imaging of the friction plane of the tested composites and a description of a typical course of material wear is described.

Supervised Machine Learning Segmentation and Quantification of Gastric Pacemaker Cells.

Interstitial Cells of Cajal (ICC) are specialized pacemaker cells that generate and actively propagate electrophysiological events called slow waves. Slow waves regulate the motility of the gastrointestinal tract necessary for digesting food. Degradation in the ICC network structure has been qualitatively associated to several gastrointestinal motility disorders. ICC network structure can be obtained using confocal microscopy, but the current limitations in imaging and segmentation techniques have hindered an accurate representation of the networks. In this study, supervised machine learning techniques were applied to extract the ICC networks from 3D confocal microscopy images. The results showed that the Fast Random Forest classification method using Trainable WEKA Segmentation outperformed the Decision Table and Naïve Bayes classification methods in sensitivity, accuracy, and F-measure. Using the Fast Random Forest classifier, 12 gastric antrum tissue blocks were segmented and variations in ICC network thickness, density and process width were quantified for the myenteric plexus ICC network (the primary pacemakers). Our findings demonstrated regional variation in ICC network density and thickness along the circumferential and longitudinal axis of the mouse antrum. An inverse relationship was observed in the distal and proximal antrum for density (proximal: 9.8±4.0% vs distal: 7.6±4.6%) and thickness (proximal: 15±3 μm vs distal: 24±10 μm). Limited variation in ICC process width was observed throughout the antrum (5±1 μm).Clinical Relevance- Detailed quantification of regional ICC structural properties will provide insights into the relationship between ICC structure, slow waves and resultant gut motility. This will improve techniques for the diagnosis and treatment of functional GI motility disorders.

Testing System for the Mechanical Properties of Small-Scale Specimens Based on 3D Microscopic Digital Image Correlation.

The testing of the mechanical properties of materials on a small scale is difficult because of the small specimen size and the difficulty of measuring the full-field strain. To tackle this problem, a testing system for investigating the mechanical properties of small-scale specimens based on the three-dimensional (3D) microscopic digital image correlation (DIC) combined with a micro tensile machine is proposed. Firstly, the testing system is described in detail, including the design of the micro tensile machine and the 3D microscopic DIC method. Then, the effects of different shape functions on the matching accuracy obtained by the inverse compositional Gauss–Newton (IC-GN) algorithm are investigated and the numerical experiment results verify that the error due to under matched shape functions is far larger than that of overmatched shape functions. The reprojection error is shown to be smaller than before when employing the modified iteratively weighted radial alignment constraint method. Both displacement and uniaxial measurements were performed to demonstrate the 3D microscopic DIC method and the testing system built. The experimental results confirm that the testing system built can accurately measure the full-field strain and mechanical properties of small-scale specimens.

DDeep3M: Docker-powered deep learning for biomedical image segmentation

BackgroundDeep learning models are turning out to be increasingly popular in biomedical image processing. The fruitful utilization of these models, in most cases, is substantially restricted by the complicated configuration of computational environments, resulting in the noteworthy increment of the time and endeavors to reproduce the outcomes of the models. New methodWe thus present a Docker-based method for better use of deep learning models and quicker reproduction of model performance for multiple data sources, permitting progressively more biomedical scientists to attempt the new technology conveniently in their domain. Here, we introduce a Docker-powered deep learning model, named as DDeep3M and validated it with the electron microscopy data volumes (microscale). ResultsDDeep3M is utilized to the 3D optical microscopy image stack in mouse brain for the image segmentation (mesoscale). It achieves high accuracy on both vessels and somata structures with all the recall/precision scores and Dice indexes over 0.96. DDeep3M also reports the state-of-the-art performance in the MRI data (macroscale) for brain tumor segmentation. Comparison with existing methodsWe compare the performance and efficiency of DDeep3M with three existing models on image datasets varying from micro- to macro-scales. ConclusionDDeep3M is a friendly, convenient and efficient tool for image segmentations in biomedical research. DDeep3M is open sourced with the codes and pretrained model weights available at https://github.com/cakuba/DDeep3m.

An integrated pipeline for high-throughput screening and profiling of spheroids using simple live image analysis of frame to frame variations.

High-throughput imaging methods can be applied to relevant cell culture models, fostering their use in research and translational applications. Improvements in microscopy, computational capabilities and data analysis have enabled high-throughput, high-content approaches from endpoint 2D microscopy images. Nonetheless, trade-offs in acquisition, computation and storage between content and throughput remain, in particular when cells and cell structures are imaged in 3D. Moreover, live 3D phase contrast microscopy images are not often amenable to analysis because of the high level of background noise. Cultures of Human induced pluripotent stem cells (hiPSC) offer unprecedented scope to profile and screen conditions affecting cell fate decisions, self-organisation and early embryonic development. However, quantifying changes in the morphology or function of cell structures derived from hiPSCs over time presents significant challenges. Here, we report a novel method based on the analysis of live phase contrast microscopy images of hiPSC spheroids. We compare self-renewing versus differentiating media conditions, which give rise to spheroids with distinct morphologies; round versus branched, respectively. These cell structures are segmented from 2D projections and analysed based on frame-to-frame variations. Importantly, a tailored convolutional neural network is trained and applied to predict culture conditions from time-frame images. We compare our results with more classic and involved endpoint 3D confocal microscopy and propose that such approaches can complement spheroid-based assays developed for the purpose of screening and profiling. This workflow can be realistically implemented in laboratories using imaging-based high-throughput methods for regenerative medicine and drug discovery.