-
[
2008]
"Public databases, such as Phenobank, provide a huge number of movies of early embryos. However, manual measurement of dynamic cell division patterns from these movies is laborious. We are developing a cell division pattern measurement system that automatically detects and tracks the nuclei in two dimensional time-lapse DIC microscope movies available from public databases. The system addresses the following problems associated with nuclear detection and tracking in public two dimensional time-lapse movies: [1] The image quality (e.g. brightness, contrast) varies among movies, which makes the nuclear detection difficult. [2] The nuclei change their shapes and undergo nuclear divisions, which increases the chance of false detection of nuclei. [3] Nuclei occasionally disappear from the movie, which makes the nuclear tracking difficult. To address the problem [1], we developed a dynamic threshold binarization method that recursively calculates a threshold for image entropy values, which distinguishes a nuclear region from cytoplasm, at each local area enclosing each nuclear region. As target movies, we sampled 104 movies of an RNAi embryo whose phenotypes was embryonic lethal on chromosome III in the Phenobank. The method detected spindles as well as nuclei at almost all the time points in all the 104 movies of various image qualities. To address the problem [2], we developed filters that remove falsely detected regions based on size, roundness, and location within an embryo. Using these filters, only 1.4% of nuclear regions detected by our system were falsely detected regions (50 movies randomly sampled from the 104 movies were examined). To address the problem [3], we utilized falsely untracked regions that have a round shape and thus most likely correspond to the nuclei. We programmed the system to restore tracking from these falsely untracked round regions. We sampled 14 movies containing falsely untracked regions. For 12 in the 14 movies, the system restored tracking from falsely untracked regions. Now, we are preparing the latest estimation of the system, and would like to report the result of the evaluation."
-
[
International Worm Meeting,
2005]
We have begun to make 4-dimensional confocal recordings of embryogenesis for a growing number of fluorescing transgenic C. elegans strains. Using a spinning-microlens confocal, 12-hour-long recordings can be acquired with stacks collected every 2.5 minutes, while maintaining the viability of the embryo. Our custom-written Worm Autoselector plug-in for ImageJ allows selection of a number of randomly oriented asynchronous embryos for individual reorientation, reconstruction, and time-annotation. The result is a pair of movies in slice4D and stereo4D QuickTimeVR format that can be played on a computer through a free stand-alone program or web browser. By reorienting and time-annotating the individual embryos, we hope to simplify the visual comparison of gene-expression patterns for many genes. Recordings from the Hunter lab, Hope lab, and Baillie/Moerman lab gene-expression profiling projects will be available for browsing by interested researchers.
-
[
International Worm Meeting,
2019]
Along with the recent advance of live-imaging technologies, enormous amounts of time-lapse microscopy images are becoming available in public databases. By applying image processing methods, quantitative data such as positions and shapes of nuclei and cells, and their temporal changes can be extracted from the images. Large-scale collections of such data are important resources for computational phenotype analysis. We therefore established a new resource of quantitative data on nuclear division dynamics in C. elegans RNAi-treated embryos by image-processing time-lapse DIC microscopy images in Phenobank. It consists of 1,579 sets of quantitative data from RNAi-treated embryos, including three sets of data for each of 518 genes that exerted defects in early embryogenesis when depleted individually by RNAi. Each data contains the nuclear regions and their temporal changes. The resource will be available in SSBD database (Tohsato et al. 2016;
http://ssbd.qbic.riken.jp). To demonstrate the usefulness of our resource for computational phenotype screening, we calculated the speed of female pronuclear migration in its RNAi-treated embryos. 12 genes showed faster or slower migration than wild-type. Out of the 12 genes, 7 genes reproduced the migration phenotype in our independent RNAi experiments. Among the 7 genes,
sds-22 and F44B9.8 exhibited remarkedly faster and slower migration respectively.
sds-22(RNAi) and F44B9.8(RNAi) embryos expressing GFP::tubulin exhibited larger and smaller microtubule sperm aster respectively, consistent with the nuclear tracking along microtubules mechanism. Convergent cross mapping (CCM) identified a causal effect of the aster size on the migration speed. Since SDS-22 down-regulates ZYG-1 kinase through GSP-1/2 in centrosomal duplications (Peel et al., 2017), we examined RNAi phenotypes of
gsp-1/2 in female pronuclear migration.
gsp-1(RNAi) exhibited faster migration, suggesting that
sds-22 regulates
zyg-1 in pronuclear migrations as well. The roles of F44B9.8, an ortholog of human replication factor C, remain unclear.These results demonstrate that our resource provides new opportunities for computational phenotype screening.
-
[
East Coast Worm Meeting,
2004]
We are developing software tools for production of 'publishable' 4D movie data from microscope recordings of GFP expression in embryos. Raw image stacks comprising a field of several randomly-oriented embryos are typically processed to produce three useful standardized data types for each embryo: sliced-4D QuickTimeVR (QTVR) movies, stereo-4D QTVR movies, and embryonic developmental expression chronograms (EDECs). A finished sliced-4D QTVR allows free animation through focus and over time for an individual embryo, re-aligned into the canonical anterior (left), posterior (right), dorsal(up), ventral(down) orientation for display of C. elegans embryos. A stereo-4D QTVR allows free rotational animation of 3D volume reconstructions (maximum point projection) around the X and Y-axes, as well as animation through time. By keying on the onset of embryonic movement, each movie isolated from an asynchronous field of embryos can be properly annotated as to developmental time. Both sliced-4D and stereo-4D QTVRs benefit from several aspects of the established QuickTimeVR data format: facile navigation of the movie using the computer mouse or keyboard, availability of free viewing software and web browser plug-ins, streaming-download web access to large movies, and flexible choices for data compression. The third data type, an EDEC, represents the full 4D data set in a 2-dimensional summary image - averaged fluorescence intensity along the one dimension of the anterior-posterior axis is plotted against time. Each of these formats has unique advantages in analysis and comparison among expression patterns for different gene products. All three should be applicable to the 'super-imposable' comparison and correlation of the expression/localization patterns of different gene products.
-
[
International Worm Meeting,
2015]
In recent years, fluorescent imaging has been used more and more to capture the state of biological systems at different moments in time. For many researchers, analysis of the fluorescent image data has become the limiting factor of this new technique. We have developed NEPIC, a semi-automated tool for finding and tracking the soma of a single neuron over an entire movie of grayscale calcium image data. When tested on calcium image movies of the AWC neuron in C. elegans under highly variant conditions, NEPIC correctly identified the neuronal soma in 95.48% of the movie frames, and successfully tracked this soma feature across 98.60% of the frame transitions. Although support for finding and tracking multiple fluorescing Regions of Interest (or fROIs) has yet to be implemented, NEPIC displays promise as a tool for assisting researchers in the bulk analysis of fluorescent imaging data.
-
[
International Worm Meeting,
2007]
A 4-D nuclear tracking system, which automatically tracks 3-dimensional (3-D) regions of nuclei of early C. elegans embryo from Nomarski DIC microscopy images, has been developed in our laboratory. The system works well on images that are conditioned so as to have a fixed image quality (e.g. brightness, contrast). It is, however, difficult for the system to process unconditioned images of different image qualities. In order to overcome this limitation of our system, we are developing software that can track nuclear regions from movies of various image qualities. Our target is tracking 2-D regions of nuclei from the fertilization to the 4-cell stage because many basic processes in development are expected to be involved in this period. In our software, local image entropy is used for detection of nuclei as in our existing system [1]. For each image the software automatically varies thresholds of local image entropy to distinguish nuclear regions from cytoplasm, whereas our existing system used a fixed threshold optimized for the conditioned images. The software first calculates a threshold for the entire image by Watsons approach [2] and determines nuclear regions using the threshold. Next, the software defines an oval area that contains each nuclear region and that is two times as large as the region, recalculates a threshold locally within the oval area by thrno method [3], and redetermines the nuclear region. This redetermination of nuclear regions is repeated 20 times for each image. This algorithm can extract regions of nuclei in the images of various image qualities. In addition to this improved nuclear detection algorithm, the software includes a forward tracking algorithm that connects a nuclear region overlapped with nuclear regions at adjacent time point, and a backward tracking algorithm that selects longer tracks of nuclear regions to exclude falsely-detected regions. As a pilot experiment, we applied the software to 27 movies of RNAi embryos in the PhenoBank database. For 26 movies, the software extracted nuclear regions well corresponding to contours of actual nuclei. For 12 in the 26 movies, the software tracked the nuclei until the 4-cell stage although sometimes there are falsely detected regions. In the remaining 14 movies, nuclei were out of focus for a long time or images leaped too much, resulting in the termination of tracking. The software will be useful not only for our 4-D nuclear tracking system but also for quantitative analysis of movies available from public databases or recorded in individual laboratories. [1] Hamahashi et al. (2005) BMC Bioinformatics. 6, 125. [2] Watson et al. (1987) Cytometry 8, 1-8. [3] Otsu. (1979) IEEE Trans. Syst. Man & Cybern. SMC-9, 62-66.
-
[
Neuronal Development, Synaptic Function, and Behavior Meeting,
2006]
Quantitative, automated analyses of movement are essential to investigations of genetic and neuronal mechanisms of behavior in C. elegans. Current tracking systems are of two main types: those that record the worm's centroid and those that record the worm's centerline. Centroid-based systems require minimal image processing but do not preserve the worm's shape, making it difficult to detect subtle changes in waveform. Centerline-based systems can detect subtle changes in waveform, but they require extensive image processing and can be confused by closed postures such as omega turns and coils. Here we report a new tracking method that reports subtle waveform effects with minimal image processing. The new method adopts the simple expedient of tracking a tiny black spot painted on the worm. The spot is a mixture of black artists' pigment and diluted Vaseline. We refer to this as the spotted worm tracking system. The system utilizes two computers. One computer commands the stage to keep the spot near the center of the optical field and records a movie of the worm. The other computer reads a pair of displacement encoders attached to the microscope stage. A virtual clap-board is used to synchronize the movie with stage-displacement data for off-line processing. The position and velocity of the spot are reconstructed from the location of the spot in successive movie frames and the position of the stage at the time of frame acquisition. Operating at 30 frames/sec, the system has a maximum spatial precision (in um) of 3.5 and 1.9 in X and Y directions, respectively. Custom software for off-line processing can distinguish between forward and reverse locomotion without user input, thereby facilitating kinetic analyses of locomotion in wild-type and mutant worms. Utilizing the system's increased spatial and temporal resolution, we have observed "micropauses," a frequent yet previously undetected behavioral state of near-zero velocity with a mean dwell time of 0.15 sec. Micropauses are observed during transitions between forward and reverse locomotion, but also during apparently continuous bouts of forward or reverse locomotion. Thus the new system may contribute to revised definitions of fundamental locomotory states in C. elegans.
-
Green, Rebecca, Ochoa, Stacy, Chow, Tiffany, Hendel, Jeff, Zhao, Zhiling, Khaliullin, Renat, Desai, Arshad, Oegema, Karen, Wang, Shaohe
[
International Worm Meeting,
2019]
An important challenge is to functionally classify the ~2000 genes (>1400 conserved) that control cell-fate specification and morphogenesis during embryogenesis. Here, we perform a 4D high-content screen by filming embryogenesis using two custom-engineered C. elegans reporter strains, following individual RNAi-based knockdown (>20,000 individual movies). We monitor (1) changes in cell fate specification, by dynamically tracking fluorescently labeled endoderm, mesoderm and ectoderm nuclei, and (2) morphogenic changes during epithelial and neuronal development by monitoring tissue position and tissue shape. Consistent and timely analysis of 20,000 movies requires automation, however, the range and complexity of 4D developmental phenotypes are not easily captured by existing automated methods. To address this challenge, we manually curated a pilot set of 500 genes (>7000 movies) and used this reference to guide the development of custom automated analysis algorithms; this effort ensured that our final automated analysis method captured observed phenotypes across a spectrum of developmental defects. For each RNAi condition, our automated analysis yields phenotypic signatures consisting of >100 continuous parameters. To evaluate the phenotypic similarity between RNAi conditions, we measure the distance between phenotypes in continuous space. To correct for the fact that a strict measure of Euclidean distance penalizes genes with more severe phenotypes, we measure the angle between the average phenotypes for the two conditions (phenotypic angle of deviation; PAD). Finally, we optimized the set of parameters used for automated comparison by assessing performance of the algorithm on a manually-annotated set of phenotypic groups. Our resulting automated method effectively identifies genes whose knockdown leads to similar phenotypes; this allows partitioning of genes into functional groups that are predicted to reflect developmental pathways and will yield a systems-level view of embryonic development. This work represents the first fully automated high-content screen of an intact developing organism and is the most complex morphological profiling effort to date.
-
[
International Worm Meeting,
2017]
Decades of research on C. elegans development have shed light on a large variety of developmental phenomena; from cell-fate decisions and transdifferentiation to cell migration and cell death. However, long-term high-resolution in-vivo imaging of these processes has been accessible only in embryos, because larvae must move and feed to develop, and existing immobilization techniques perturb development and compromise animal viability. To overcome this problem, we have developed a microfluidic setup to simultaneously follow development of ten C. elegans larvae at high spatiotemporal resolution from hatching to adulthood (~3 days). Animals grown in microchambers are periodically immobilized by compression to allow high-quality imaging of even weak fluorescence signals. We show that this technology can be used to follow cell-fate decisions, cell death programs, and transdifferentiation events with Nomarski and multichannel fluorescence microscopy. We also obtain cell-cycle timing statistics during the formation of the C. elegans vulva in a large number of animals, exposing the limits of fidelity and robustness of C. elegans development. Finally, we develop algorithms for automated image registration to generate time-lapse movies. For the first time, these movies enable us to visualize and quantify highly-diverse processes such as neural arborization or cell divisions during gonadogenesis in a feeding, moving, and growing animal. Our technique opens the door to quantitative analysis of time-dependent phenomena governing cellular behavior during C. elegans larval development.
-
Kim, J., Murray, J., Packer, J., Huynh, C., Zhu, Q, Waterston, R.
[
International Worm Meeting,
2019]
Using single cell RNA-seq methods we profiled the transcripts of 84,625 cells from different C. elegans embryo populations. Exploiting known expression markers and 3D confocal movies from the EPiC2 database we identify 105 distinct terminal cell types and almost all preterminal cell types from the 50-cell stage on. Using the expression profiles of these annotated cells, we find that lineage signatures of AB progeny in the early embryo break down in the final two cell divisions. Also, distinct lineages that produce the same terminal anatomic cell type converge to a homogeneous transcriptomic state. Finally, we find evidence of multi-lineage priming, where expression of transcript factors in parent cells are selectively maintained in only one daughter cell.