The Cell Tracking Challenge: 10 years of objective benchmarking.

The Cell Tracking Challenge is an ongoing benchmarking initiative that has become a reference in cell segmentation and tracking algorithm development. Here, we present a significant number of improvements introduced in the challenge since our 2017 report. These include the creation of a new segmentation-only benchmark, the enrichment of the dataset repository with new datasets that increase its diversity and complexity, and the creation of a silver standard reference corpus based on the most competitive results, which will be of particular interest for data-hungry deep learning-based strategies. Furthermore, we present the up-to-date cell segmentation and tracking leaderboards, an in-depth analysis of the relationship between the performance of the state-of-the-art methods and the properties of the datasets and annotations, and two novel, insightful studies about the generalizability and the reusability of top-performing methods. These studies provide critical practical conclusions for both developers and users of traditional and machine learning-based cell segmentation and tracking algorithms.

Phosphorylation-Dependent Feedback Inhibition of RIG-I by DAPK1 Identified by Kinome-wide siRNA Screening.

Cell-autonomous induction of type I interferon must be stringently regulated. Rapid induction is key to control virus infection, whereas proper limitation of signaling is essential to prevent immunopathology and autoimmune disease. Using unbiased kinome-wide RNAi screening followed by thorough validation, we identified 22 factors that regulate RIG-I/IRF3 signaling activity. We describe a negative-feedback mechanism targeting RIG-I activity, which is mediated by death associated protein kinase 1 (DAPK1). RIG-I signaling triggers DAPK1 kinase activation, and active DAPK1 potently inhibits RIG-I stimulated IRF3 activity and interferon-beta production. DAPK1 phosphorylates RIG-I in vitro at previously reported as well as other sites that limit 5'ppp-dsRNA sensing and virtually abrogate RIG-I activation.

Replication of ribosomal DNA in Arabidopsis occurs both inside and outside the nucleolus during S phase progression.

Ribosomal RNA genes (rDNA) have been used as valuable experimental systems in numerous studies. Here, we focus on elucidating the spatiotemporal organisation of rDNA replication in Arabidopsis thaliana To determine the subnuclear distribution of rDNA and the progression of its replication during the S phase, we apply 5-ethynyl-2'-deoxyuridine (EdU) labelling, fluorescence-activated cell sorting, fluorescence in situ hybridization and structured illumination microscopy. We show that rDNA is replicated inside and outside the nucleolus, where active transcription occurs at the same time. Nascent rDNA shows a maximum of nucleolar associations during early S phase. In addition to EdU patterns typical for early or late S phase, we describe two intermediate EdU profiles characteristic for mid S phase. Moreover, the use of lines containing mutations in the chromatin assembly factor-1 gene fas1 and wild-type progeny of fas1xfas2 crosses depleted of inactive copies allows for selective observation of the replication pattern of active rDNA. High-resolution data are presented, revealing the culmination of replication in the mid S phase in the nucleolus and its vicinity. Taken together, our results provide a detailed snapshot of replication of active and inactive rDNA during S phase progression.

An objective comparison of cell-tracking algorithms.

We present a combined report on the results of three editions of the Cell Tracking Challenge, an ongoing initiative aimed at promoting the development and objective evaluation of cell segmentation and tracking algorithms. With 21 participating algorithms and a data repository consisting of 13 data sets from various microscopy modalities, the challenge displays today's state-of-the-art methodology in the field. We analyzed the challenge results using performance measures for segmentation and tracking that rank all participating methods. We also analyzed the performance of all of the algorithms in terms of biological measures and practical usability. Although some methods scored high in all technical aspects, none obtained fully correct solutions. We found that methods that either take prior information into account using learning strategies or analyze cells in a global spatiotemporal video context performed better than other methods under the segmentation and tracking scenarios included in the challenge.

The SUMOylation Pathway Restricts Gene Transduction by Adeno-Associated Viruses.

Adeno-associated viruses are members of the genus dependoviruses of the parvoviridae family. AAV vectors are considered promising vectors for gene therapy and genetic vaccination as they can be easily produced, are highly stable and non-pathogenic. Nevertheless, transduction of cells in vitro and in vivo by AAV in the absence of a helper virus is comparatively inefficient requiring high multiplicity of infection. Several bottlenecks for AAV transduction have previously been described, including release from endosomes, nuclear transport and conversion of the single stranded DNA into a double stranded molecule. We hypothesized that the bottlenecks in AAV transduction are, in part, due to the presence of host cell restriction factors acting directly or indirectly on the AAV-mediated gene transduction. In order to identify such factors we performed a whole genome siRNA screen which identified a number of putative genes interfering with AAV gene transduction. A number of factors, yielding the highest scores, were identified as members of the SUMOylation pathway. We identified Ubc9, the E2 conjugating enzyme as well as Sae1 and Sae2, enzymes responsible for activating E1, as factors involved in restricting AAV. The restriction effect, mediated by these factors, was validated and reproduced independently. Our data indicate that SUMOylation targets entry of AAV capsids and not downstream processes of uncoating, including DNA single strand conversion or DNA damage signaling. We suggest that transiently targeting SUMOylation will enhance application of AAV in vitro and in vivo.

Fibroblast growth factor and canonical WNT/ß-catenin signaling cooperate in suppression of chondrocyte differentiation in experimental models of FGFR signaling in cartilage.

Aberrant fibroblast growth factor (FGF) signaling disturbs chondrocyte differentiation in skeletal dysplasia, but the mechanisms underlying this process remain unclear. Recently, FGF was found to activate canonical WNT/ß-catenin pathway in chondrocytes via Erk MAP kinase-mediated phosphorylation of WNT co-receptor Lrp6. Here, we explore the cellular consequences of such a signaling interaction. WNT enhanced the FGF-mediated suppression of chondrocyte differentiation in mouse limb bud micromass and limb organ cultures, leading to inhibition of cartilage nodule formation in micromass cultures, and suppression of growth in cultured limbs. Simultaneous activation of the FGF and WNT/ß-catenin pathways resulted in loss of chondrocyte extracellular matrix, expression of genes typical for mineralized tissues and alteration of cellular shape. WNT enhanced the FGF-mediated downregulation of chondrocyte proteoglycan and collagen extracellular matrix via inhibition of matrix synthesis and induction of proteinases involved in matrix degradation. Expression of genes regulating RhoA GTPase pathway was induced by FGF in cooperation with WNT, and inhibition of the RhoA signaling rescued the FGF/WNT-mediated changes in chondrocyte cellular shape. Our results suggest that aberrant FGF signaling cooperates with WNT/ß-catenin in suppression of chondrocyte differentiation.

A benchmark for comparison of cell tracking algorithms.

MOTIVATION: Automatic tracking of cells in multidimensional time-lapse fluorescence microscopy is an important task in many biomedical applications. A novel framework for objective evaluation of cell tracking algorithms has been established under the auspices of the IEEE International Symposium on Biomedical Imaging 2013 Cell Tracking Challenge. In this article, we present the logistics, datasets, methods and results of the challenge and lay down the principles for future uses of this benchmark. RESULTS: The main contributions of the challenge include the creation of a comprehensive video dataset repository and the definition of objective measures for comparison and ranking of the algorithms. With this benchmark, six algorithms covering a variety of segmentation and tracking paradigms have been compared and ranked based on their performance on both synthetic and real datasets. Given the diversity of the datasets, we do not declare a single winner of the challenge. Instead, we present and discuss the results for each individual dataset separately. AVAILABILITY AND IMPLEMENTATION: The challenge Web site (http://www.codesolorzano.com/celltrackingchallenge) provides access to the training and competition datasets, along with the ground truth of the training videos. It also provides access to Windows and Linux executable files of the evaluation software and most of the algorithms that competed in the challenge.

Performance and sensitivity evaluation of 3D spot detection methods in confocal microscopy.

Reliable 3D detection of diffraction-limited spots in fluorescence microscopy images is an important task in subcellular observation. Generally, fluorescence microscopy images are heavily degraded by noise and non-specifically stained background, making reliable detection a challenging task. In this work, we have studied the performance and parameter sensitivity of eight recent methods for 3D spot detection. The study is based on both 3D synthetic image data and 3D real confocal microscopy images. The synthetic images were generated using a simulator modeling the complete imaging setup, including the optical path as well as the image acquisition process. We studied the detection performance and parameter sensitivity under different noise levels and under the influence of uneven background signal. To evaluate the parameter sensitivity, we propose a novel measure based on the gradient magnitude of the F1 score. We measured the success rate of the individual methods for different types of the image data and found that the type of image degradation is an important factor. Using the F1 score and the newly proposed sensitivity measure, we found that the parameter sensitivity is not necessarily proportional to the success rate of a method. This also provided an explanation why the best performing method for synthetic data was outperformed by other methods when applied to the real microscopy images. On the basis of the results obtained, we conclude with the recommendation of the HDome method for data with relatively low variations in quality, or the Sorokin method for image sets in which the quality varies more. We also provide alternative recommendations for high-quality images, and for situations in which detailed parameter tuning might be deemed expensive.

Live-cell assays to identify regulators of ER-to-Golgi trafficking.

We applied fluorescence microscopy-based quantitative assays to living cells to identify regulators of endoplasmic reticulum (ER)-to-Golgi trafficking and/or Golgi complex maintenance. We first validated an automated procedure to identify factors which influence Golgi-to-ER relocalization of GalT-CFP (ß1,4-galactosyltransferase I-cyan fluorescent protein) after brefeldin A (BFA) addition and/or wash-out. We then tested 14 proteins that localize to the ER and/or Golgi complex when overexpressed for a role in ER-to-Golgi trafficking. Nine of them interfered with the rate of BFA-induced redistribution of GalT-CFP from the Golgi complex to the ER, six of them interfered with GalT-CFP redistribution from the ER to a juxtanuclear region (i.e. the Golgi complex) after BFA wash-out and six of them were positive effectors in both assays. Notably, our live-cell approach captures regulator function in ER-to-Golgi trafficking, which was missed in previous fixed cell assays, as well as assigns putative roles for other less characterized proteins. Moreover, we show that our assays can be extended to RNAi and chemical screens.

Normalizing for individual cell population context in the analysis of high-content cellular screens.

BACKGROUND: High-content, high-throughput RNA interference (RNAi) offers unprecedented possibilities to elucidate gene function and involvement in biological processes. Microscopy based screening allows phenotypic observations at the level of individual cells. It was recently shown that a cell's population context significantly influences results. However, standard analysis methods for cellular screens do not currently take individual cell data into account unless this is important for the phenotype of interest, i.e. when studying cell morphology. RESULTS: We present a method that normalizes and statistically scores microscopy based RNAi screens, exploiting individual cell information of hundreds of cells per knockdown. Each cell's individual population context is employed in normalization. We present results on two infection screens for hepatitis C and dengue virus, both showing considerable effects on observed phenotypes due to population context. In addition, we show on a non-virus screen that these effects can be found also in RNAi data in the absence of any virus. Using our approach to normalize against these effects we achieve improved performance in comparison to an analysis without this normalization and hit scoring strategy. Furthermore, our approach results in the identification of considerably more significantly enriched pathways in hepatitis C virus replication than using a standard analysis approach. CONCLUSIONS: Using a cell-based analysis and normalization for population context, we achieve improved sensitivity and specificity not only on a individual protein level, but especially also on a pathway level. This leads to the identification of new host dependency factors of the hepatitis C and dengue viruses and higher reproducibility of results.

Detecting host factors involved in virus infection by observing the clustering of infected cells in siRNA screening images.

MOTIVATION: Detecting human proteins that are involved in virus entry and replication is facilitated by modern high-throughput RNAi screening technology. However, hit lists from different laboratories have shown only little consistency. This may be caused by not only experimental discrepancies, but also not fully explored possibilities of the data analysis. We wanted to improve reliability of such screens by combining a population analysis of infected cells with an established dye intensity readout. RESULTS: Viral infection is mainly spread by cell-cell contacts and clustering of infected cells can be observed during spreading of the infection in situ and in vivo. We employed this clustering feature to define knockdowns which harm viral infection efficiency of human Hepatitis C Virus. Images of knocked down cells for 719 human kinase genes were analyzed with an established point pattern analysis method (Ripley's K-function) to detect knockdowns in which virally infected cells did not show any clustering and therefore were hindered to spread their infection to their neighboring cells. The results were compared with a statistical analysis using a common intensity readout of the GFP-expressing viruses and a luciferase-based secondary screen yielding five promising host factors which may suit as potential targets for drug therapy. CONCLUSION: We report of an alternative method for high-throughput imaging methods to detect host factors being relevant for the infection efficiency of viruses. The method is generic and has the potential to be used for a large variety of different viruses and treatments being screened by imaging techniques.

Single-cell-based image analysis of high-throughput cell array screens for quantification of viral infection.

The identification of eukaryotic genes involved in virus entry and replication is important for understanding viral infection. Our goal is to develop a siRNA-based screening system using cell arrays and high-throughput (HT) fluorescence microscopy. A central issue is efficient, robust, and automated single-cell-based analysis of massive image datasets. We have developed an image analysis approach that comprises (i) a novel, gradient-based thresholding scheme for cell nuclei segmentation which does not require subsequent postprocessing steps for separation of clustered nuclei, (ii) quantification of the virus signal in the neighborhood of cell nuclei, (iii) localization of regions with transfected cells by combining model-based circle fitting and grid fitting, (iv) cell classification as infected or noninfected, and (v) image quality control (e.g., identification of out-of-focus images). We compared the results of our nucleus segmentation approach with a previously developed scheme of adaptive thresholding with subsequent separation of nuclear clusters. Our approach, which does not require a postprocessing step for the separation of nuclear clusters, correctly segmented 97.1% of the nuclei, whereas the previous scheme achieved 95.8%. Using our algorithm for the detection of out-of-focus images, we obtained a high discrimination power of 99.4%. Our overall approach has been applied to more than 55,000 images of cells infected by either hepatitis C or dengue virus. Reduced infection rates were correctly detected in positive siRNA controls, as well as for siRNAs targeting, for example, cellular genes involved in viral infection. Our image analysis approach allows for the automatic and accurate determination of changes in viral infection based on high-throughput single-cell-based siRNA cell array imaging experiments.

Reciprocal Effects of Fibroblast Growth Factor Receptor Signaling on Dengue Virus Replication and Virion Production.

Dengue virus (DENV) is a human arboviral pathogen accounting for 390 million infections every year. The available vaccine has limited efficacy, and DENV-specific drugs have not been generated. To better understand DENV-host cell interaction, we employed RNA interference-based screening of the human kinome and identified fibroblast growth factor receptor 4 (FGFR4) to control the DENV replication cycle. Pharmacological inhibition of FGFR exerts a reciprocal effect by reducing DENV RNA replication and promoting the production of infectious virus particles. Addressing the latter effect, we found that the FGFR signaling pathway modulates intracellular distribution of DENV particles in a PI3K-dependent manner. Upon FGFR inhibition, virions accumulate in the trans-Golgi network compartment, where they undergo enhanced maturation cleavage of the envelope protein precursor membrane (prM), rendering virus particles more infectious. This study reveals an unexpected reciprocal role of a cellular receptor tyrosine kinase regulating DENV RNA replication and the production of infectious virions.

Reprogramming of Adult Peripheral Blood Cells into Human Induced Pluripotent Stem Cells as a Safe and Accessible Source of Endothelial Cells.

New approaches in regenerative medicine and vasculogenesis have generated a demand for sufficient numbers of human endothelial cells (ECs). ECs and their progenitors reside on the interior surface of blood and lymphatic vessels or circulate in peripheral blood; however, their numbers are limited, and they are difficult to expand after isolation. Recent advances in human induced pluripotent stem cell (hiPSC) research have opened possible avenues to generate unlimited numbers of ECs from easily accessible cell sources, such as the peripheral blood. In this study, we reprogrammed peripheral blood mononuclear cells, human umbilical vein endothelial cells (HUVECs), and human saphenous vein endothelial cells (HSVECs) into hiPSCs and differentiated them into ECs. The phenotype profiles, functionality, and genome stability of all hiPSC-derived ECs were assessed and compared with HUVECs and HSVECs. hiPSC-derived ECs resembled their natural EC counterparts, as shown by the expression of the endothelial surface markers CD31 and CD144 and the results of the functional analysis. Higher expression of endothelial progenitor markers CD34 and kinase insert domain receptor (KDR) was measured in hiPSC-derived ECs. An analysis of phosphorylated histone H2AX (γH2AX) foci revealed that an increased number of DNA double-strand breaks upon reprogramming into pluripotent cells. However, differentiation into ECs restored a normal number of γH2AX foci. Our hiPSCs retained a normal karyotype, with the exception of the HSVEC-derived hiPSC line, which displayed mosaicism due to a gain of chromosome 1. Peripheral blood from adult donors is a suitable source for the unlimited production of patient-specific ECs through the hiPSC interstage. hiPSC-derived ECs are fully functional and comparable to natural ECs. The protocol is eligible for clinical applications in regenerative medicine, if the genomic stability of the pluripotent cell stage is closely monitored.

Fast point-based 3-D alignment of live cells.

Typical time intervals between acquisitions of three-dimensional (3-D) images of the same cell in live cell imaging are in the orders of minutes. In the meantime, the live cell can move in a water basin on the stage. This movement can hamper the studies of intranuclear processes. We propose a fast point-based image registration method for the suppression of the movement of a cell as a whole in the image data. First, centroids of certain intracellular objects are computed for each image in a time-lapse series. Then, a matching between the centroids, which have the maximal number of pairs, is sought between consecutive point sets by a 3-D extension of a two-dimensional fast point pattern matching method, which is invariant to rotation, translation, local distortion, and extra/missing points. The proposed 3-D extension assumes rotations only around the z axis to retain the complexity of the original method. The final step involves computing the optimal fully 3-D transformation between images from corresponding points in the least-squares manner. The robustness of the method was evaluated on generated data. The results of the simulations show that the method is very precise and its correctness can be estimated. This article also presents two practical application examples, namely the registration of images of HP1 domains and the registration of images of telomeres. More than 97% of time-consecutive images were successfully registered. The results show that the method is very well suited to live cell imaging.

Cell Tracking Accuracy Measurement Based on Comparison of Acyclic Oriented Graphs.

Tracking motile cells in time-lapse series is challenging and is required in many biomedical applications. Cell tracks can be mathematically represented as acyclic oriented graphs. Their vertices describe the spatio-temporal locations of individual cells, whereas the edges represent temporal relationships between them. Such a representation maintains the knowledge of all important cellular events within a captured field of view, such as migration, division, death, and transit through the field of view. The increasing number of cell tracking algorithms calls for comparison of their performance. However, the lack of a standardized cell tracking accuracy measure makes the comparison impracticable. This paper defines and evaluates an accuracy measure for objective and systematic benchmarking of cell tracking algorithms. The measure assumes the existence of a ground-truth reference, and assesses how difficult it is to transform a computed graph into the reference one. The difficulty is measured as a weighted sum of the lowest number of graph operations, such as split, delete, and add a vertex and delete, add, and alter the semantics of an edge, needed to make the graphs identical. The measure behavior is extensively analyzed based on the tracking results provided by the participants of the first Cell Tracking Challenge hosted by the 2013 IEEE International Symposium on Biomedical Imaging. We demonstrate the robustness and stability of the measure against small changes in the choice of weights for diverse cell tracking algorithms and fluorescence microscopy datasets. As the measure penalizes all possible errors in the tracking results and is easy to compute, it may especially help developers and analysts to tune their algorithms according to their needs.

Automated acquisition and processing of multidimensional image data in confocal in vivo microscopy.

The successful development of visualization techniques for live cell imaging leads to the development of suitable software for the acquisition and processing of multidimensional image data. This report compares several possible approaches to image acquisition and processing in confocal in vivo microscopy and suggests new alternatives to the published methods. Special attention is paid to spinning disk systems based either on a classical Nipkow disk or on the microlens principle. This study shows how to optimize image acquisition process in live cell studies using camera binning feature and how to perform object tracking using a new fast image registration method based on the graph theory.

miR-17-5p regulates endocytic trafficking through targeting TBC1D2/Armus.

miRNA cluster miR-17-92 is known as oncomir-1 due to its potent oncogenic function. miR-17-92 is a polycistronic cluster that encodes 6 miRNAs, and can both facilitate and inhibit cell proliferation. Known targets of miRNAs encoded by this cluster are largely regulators of cell cycle progression and apoptosis. Here, we show that miRNAs encoded by this cluster and sharing the seed sequence of miR-17 exert their influence on one of the most essential cellular processes - endocytic trafficking. By mRNA expression analysis we identified that regulation of endocytic trafficking by miR-17 can potentially be achieved by targeting of a number of trafficking regulators. We have thoroughly validated TBC1D2/Armus, a GAP of Rab7 GTPase, as a novel target of miR-17. Our study reveals regulation of endocytic trafficking as a novel function of miR-17, which might act cooperatively with other functions of miR-17 and related miRNAs in health and disease.

Micro axial tomography: a miniaturized, versatile stage device to overcome resolution anisotropy in fluorescence light microscopy.

With the development of novel fluorescence techniques, high resolution light microscopy has become a challenging technique for investigations of the three-dimensional (3D) micro-cosmos in cells and sub-cellular components. So far, all fluorescence microscopes applied for 3D imaging in biosciences show a spatially anisotropic point spread function resulting in an anisotropic optical resolution or point localization precision. To overcome this shortcoming, micro axial tomography was suggested which allows object tilting on the microscopic stage and leads to an improvement in localization precision and spatial resolution. Here, we present a miniaturized device which can be implemented in a motor driven microscope stage. The footprint of this device corresponds to a standard microscope slide. A special glass fiber can manually be adjusted in the object space of the microscope lens. A stepwise fiber rotation can be controlled by a miniaturized stepping motor incorporated into the device. By means of a special mounting device, test particles were fixed onto glass fibers, optically localized with high precision, and automatically rotated to obtain views from different perspective angles under which distances of corresponding pairs of objects were determined. From these angle dependent distance values, the real 3D distance was calculated with a precision in the ten nanometer range (corresponding here to an optical resolution of 10-30 nm) using standard microscopic equipment. As a proof of concept, the spindle apparatus of a mature mouse oocyte was imaged during metaphase II meiotic arrest under different perspectives. Only very few images registered under different rotation angles are sufficient for full 3D reconstruction. The results indicate the principal advantage of the micro axial tomography approach for many microscopic setups therein and also those of improved resolutions as obtained by high precision localization determination.

Recruitment and activation of a lipid kinase by hepatitis C virus NS5A is essential for integrity of the membranous replication compartment.

Hepatitis C virus (HCV) is a major causative agent of chronic liver disease in humans. To gain insight into host factor requirements for HCV replication, we performed a siRNA screen of the human kinome and identified 13 different kinases, including phosphatidylinositol-4 kinase III alpha (PI4KIIIα), as being required for HCV replication. Consistent with elevated levels of the PI4KIIIα product phosphatidylinositol-4-phosphate (PI4P) detected in HCV-infected cultured hepatocytes and liver tissue from chronic hepatitis C patients, the enzymatic activity of PI4KIIIα was critical for HCV replication. Viral nonstructural protein 5A (NS5A) was found to interact with PI4KIIIα and stimulate its kinase activity. The absence of PI4KIIIα activity induced a dramatic change in the ultrastructural morphology of the membranous HCV replication complex. Our analysis suggests that the direct activation of a lipid kinase by HCV NS5A contributes critically to the integrity of the membranous viral replication complex.