FogBank: a single cell segmentation across multiple cell lines and image modalities

Background

Many cell lines currently used in medical research, such as cancer cells or stem cells, grow in confluent sheets or colonies. The biology of individual cells provide valuable information, thus the separation of touching cells in these microscopy images is critical for counting, identification and measurement of individual cells. Over-segmentation of single cells continues to be a major problem for methods based on morphological watershed due to the high level of noise in microscopy cell images. There is a need for a new segmentation method that is robust over a wide variety of biological images and can accurately separate individual cells even in challenging datasets such as confluent sheets or colonies.

Results

We present a new automated segmentation method called FogBank that accurately separates cells when confluent and touching each other. This technique is successfully applied to phase contrast, bright field, fluorescence microscopy and binary images. The method is based on morphological watershed principles with two new features to improve accuracy and minimize over-segmentation.First, FogBank uses histogram binning to quantize pixel intensities which minimizes the image noise that causes over-segmentation. Second, FogBank uses a geodesic distance mask derived from raw images to detect the shapes of individual cells, in contrast to the more linear cell edges that other watershed-like algorithms produce.We evaluated the segmentation accuracy against manually segmented datasets using two metrics. FogBank achieved segmentation accuracy on the order of 0.75 (1 being a perfect match). We compared our method with other available segmentation techniques in term of achieved performance over the reference data sets. FogBank outperformed all related algorithms. The accuracy has also been visually verified on data sets with 14 cell lines across 3 imaging modalities leading to 876 segmentation evaluation images.

Conclusions

FogBank produces single cell segmentation from confluent cell sheets with high accuracy. It can be applied to microscopy images of multiple cell lines and a variety of imaging modalities. The code for the segmentation method is available as open-source and includes a Graphical User Interface for user friendly execution.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-014-0431-x) contains supplementary material, which is available to authorized users.     

A semi-automated assessment of cell size and shape in monolayers,with optional adjustment for the cell-cell border width-application to human corneal endothelium

Measurements of large numbers of feature sizes within defined domains (e.g. cell areas within cell-cell borders) can be a time-consuming activity, but automation that includes defining such domains has not be proven to be very reliable. Other alternatives are therefore needed, and the goal of the present studies was both to develop a semi-automated (interactive) measurement system for cell areas and to carefully compare the output to that obtained using a manual digitiser pad method. A particular interest was in the contribution made by the cell-cell border zones. Non-contact specular micrographs of central corneal endothelium were obtained from 20 white male adults, aged 40-60 years. An overlay of the endothelial image was generated manually, from which the areas of around 200 cells were measured manually with a digitiser pad and also by a computer-assisted scanning method. The pad data was typed into a spread sheet along with details of the number of cell apices (sides). The computerised analysis identified borders of the same cells on the overlay, reduced these borders to a minimum, and then assessed cell area by the pixel count along with the number of neighbouring cells (to give cell sides data). The average cell area was 393 +/- 28 and 422 +/- 29 microm(2) (mean+/-SD) by the digitiser pad and computer-based methods, respectively. The average areas for each cell type were 153, 270, 392, 519 and 685 microm(2) for 4-, 5-, 6-, 7- and 8-sided cells, respectively. Assessment of the relationship between cell area and the number of cell sides (area-side relationships) showed a highly significant and positive correlation (P<0.001; r(2)=0.865). Comparing the two methods, the average cell area was 7.5% higher in the computer scan method, and this is attributed to the fact that the contribution made by the cell borders (the para-cellular space) had been essentially eliminated. A proportional correction factor can be applied to add back the cell borders/intercellular space to the computerised output, and examples are given based on using the average data from digitiser pad for each cell type. In conclusion, a computer assisted method has been developed to simultaneously provide data on the variance in cell areas (polymegethism) and cell shape (pleomorphism) from overlays of 200 cells from human corneal endothelial images, with the cell border zone corrected to allow for a finite para-cellular space.     

An adjustable-threshold algorithm for the identification of objects in three-dimensional images

MOTIVATION:To develop a highly accurate, practical and fast automated segmentation algorithm for three-dimensional images containing biological objects. To test the algorithm on images of the Drosophila brain, and identify, count and determine the locations of neurons in the images. RESULTS: A new adjustable-threshold algorithm was developed to efficiently segment fluorescently labeled objects contained within three-dimensional images obtained from laser scanning confocal microscopy, or two-photon microscopy. The result of the test segmentation with Drosophila brain images showed that the algorithm is extremely accurate and provided detailed information about the locations of neurons in the Drosophila brain. Centroids of each object (nucleus of each neuron) were also recorded into an algebraic matrix that describes the locations of the neurons. AVAILABILITY: Interested parties should send their request for the NeuronMapper(TM) program with the segmentation algorithm to artemp@bcm.tmc.edu.     

Cell Geometry Guides the Dynamic Targeting of Apoplastic GPI-Linked Lipid Transfer Protein to Cell Wall Elements and Cell Borders in Arabidopsis thaliana

During cellular morphogenesis, changes in cell shape and cell junction topology are fundamental to normal tissue and organ development. Here we show that apoplastic Glycophosphatidylinositol (GPI)-anchored Lipid Transfer Protein (LTPG) is excluded from cell junctions and flat wall regions, and passively accumulates around their borders in the epidermal cells of Arabidopsis thaliana. Beginning with intense accumulation beneath highly curved cell junction borders, this enrichment is gradually lost as cells become more bulbous during their differentiation. In fully mature epidermal cells, YFP-LTPG often shows a fibrous cellulose microfibril-like pattern within the bulging outer faces. Physical contact between a flat glass surface and bulbous cell surface induces rapid and reversible evacuation from contact sites and accumulation to the curved wall regions surrounding the contact borders. Thus, LTPG distribution is dynamic, responding to changes in cell shape and wall curvature during cell growth and differentiation. We hypothesize that this geometry-based mechanism guides wax-carrying LTPG to functional sites, where it may act to “seal” the vulnerable border surrounding cell-cell junctions and assist in cell wall fortification and cuticular wax deposition.     

PombeX: Robust Cell Segmentation for Fission Yeast Transillumination Images

Schizosaccharomyces pombe shares many genes and proteins with humans and is a good model for chromosome behavior and DNA dynamics, which can be analyzed by visualizing the behavior of fluorescently tagged proteins in vivo. Performing a genome-wide screen for changes in such proteins requires developing methods that automate analysis of a large amount of images, the first step of which requires robust segmentation of the cell. We developed a segmentation system, PombeX, that can segment cells from transmitted illumination images with focus gradient and varying contrast. Corrections for focus gradient are applied to the image to aid in accurate detection of cell membrane and cytoplasm pixels, which is used to generate initial contours for cells. Gradient vector flow snake evolution is used to obtain the final cell contours. Finally, a machine learning-based validation of cell contours removes most incorrect or spurious contours. Quantitative evaluations show overall good segmentation performance on a large set of images, regardless of differences in image quality, lighting condition, focus condition and phenotypic profile. Comparisons with recent related methods for yeast cells show that PombeX outperforms current methods, both in terms of segmentation accuracy and computational speed.     

The application of convolution neural network based cell segmentation during cryopreservation

For most of the cells, water permeability and plasma membrane properties play a vital role in the optimal protocol for successful cryopreservation. Measuring the water permeability of cells during subzero temperature is essential. So far, there is no perfect segmentation technique to be used for the image processing task on subzero temperature accurately. The ice formation and variable background during freezing posed a significant challenge for most of the conventional segmentation algorithms. Thus, a robust and accurate segmentation approach that can accurately extract cells from extracellular ice that surrounding the cell boundary is needed. Therefore, we propose a convolutional neural network (CNN) architecture similar to U-Net but differs from those conventionally used in computer vision to extract all the cell boundaries as they shrank in the engulfing ice. The images used was obtained from the cryo-stage microscope, and the data was validated using the Hausdorff distance, means ± standard deviation for different methods of segmentation result using the CNN model. The experimental results prove that the typical CNN model extracts cell borders contour from the background in its subzero state more coherent and effective as compared to other traditional segmentation approaches.     

Segmentation of the Clustered Cells with Optimized Boundary Detection in Negative Phase Contrast Images

Cell image segmentation plays a central role in numerous biology studies and clinical applications. As a result, the development of cell image segmentation algorithms with high robustness and accuracy is attracting more and more attention. In this study, an automated cell image segmentation algorithm is developed to get improved cell image segmentation with respect to cell boundary detection and segmentation of the clustered cells for all cells in the field of view in negative phase contrast images. A new method which combines the thresholding method and edge based active contour method was proposed to optimize cell boundary detection. In order to segment clustered cells, the geographic peaks of cell light intensity were utilized to detect numbers and locations of the clustered cells. In this paper, the working principles of the algorithms are described. The influence of parameters in cell boundary detection and the selection of the threshold value on the final segmentation results are investigated. At last, the proposed algorithm is applied to the negative phase contrast images from different experiments. The performance of the proposed method is evaluated. Results show that the proposed method can achieve optimized cell boundary detection and highly accurate segmentation for clustered cells.     

Cell cycle progression is required for zebrafish somite morphogenesis but not segmentation clock function

Cell division, differentiation and morphogenesis are coordinated during embryonic development, and frequently are in disarray in pathologies such as cancer. Here, we present a zebrafish mutant that ceases mitosis at the beginning of gastrulation, but that undergoes axis elongation and develops blood, muscle and a beating heart. We identify the mutation as being in early mitotic inhibitor 1 (emi1), a negative regulator of the Anaphase Promoting Complex, and use the mutant to examine the role of the cell cycle in somitogenesis. The mutant phenotype indicates that axis elongation during the segmentation period is driven substantially by cell migration. We find that the segmentation clock, which regulates somitogenesis, functions normally in the absence of cell cycle progression, and observe that mitosis is a modest source of noise for the clock. Somite morphogenesis involves the epithelialization of the somite border cells around a core of mesenchyme. As in wild-type embryos, somite boundary cells are polarized along a Fibronectin matrix in emi1(-/-). The mutants also display evidence of segment polarity. However, in the absence of a normal cell cycle, somites appear to hyper-epithelialize, as the internal mesenchymal cells exit the core of the somite after initial boundary formation. Thus, cell cycle progression is not required during the segmentation period for segmentation clock function but is necessary for the normal segmental arrangement of epithelial borders and internal mesenchymal cells.     

Object Segmentation and Ground Truth in 3D Embryonic Imaging

Many questions in developmental biology depend on measuring the position and movement of individual cells within developing embryos. Yet, tools that provide this data are often challenged by high cell density and their accuracy is difficult to measure. Here, we present a three-step procedure to address this problem. Step one is a novel segmentation algorithm based on image derivatives that, in combination with selective post-processing, reliably and automatically segments cell nuclei from images of densely packed tissue. Step two is a quantitative validation using synthetic images to ascertain the efficiency of the algorithm with respect to signal-to-noise ratio and object density. Finally, we propose an original method to generate reliable and experimentally faithful ground truth datasets: Sparse-dense dual-labeled embryo chimeras are used to unambiguously measure segmentation errors within experimental data. Together, the three steps outlined here establish a robust, iterative procedure to fine-tune image analysis algorithms and microscopy settings associated with embryonic 3D image data sets.     

Spatial regulation of DELTA expression mediates NOTCH signalling for segmentation of Drosophila legs.

The Notch (N) signalling pathway is recruited for segregation of cell fates in a number of Drosophila tissue types. We show here that N dependent segmentation of Drosophila legs is regulated by a dynamic pattern of expression of its ligand, DELTA (DL). During third larval instar and early stages of pupation, high levels of DL expression is seen in stripes of cells in the leg imaginal discs which later form the proximal borders of leg joints. These domains also displayed heightened Dl enhancer activity. During subsequent stages of pupation, following segmentation of the leg primordium, DL expression becomes uniform throughout these segments barring the joints. We further show that regulatory Dl mutations or mis-expression of DL abolish leg segmentation. Domains of N signalling for segmentation of legs of flies are thus set up by a stringent spatial regulation of expression of its ligand at the segment border. Further, a comparable role of DL in antennal development reveals a common paradigm of DL-N signalling for segmentation of appendages in flies.     

Deformable models for segmentation of CLSM tissue images and its application in FISH signal analysis.

In this paper we present an application of deformable models for the segmentation of volumetric tissue images. The three-dimensional images are obtained using confocal microscope. The segmented images have been used for the quantitative analysis of the Fluorescence In Situ Hybridization (FISH) signals. An ellipsoidal surface initialized around the cell of interest acts as a deformable model. The deformable model surface voxels are subjected to various internal and external forces derived from underlying image features as well as externally imposed constraints. The deformable model converges to the optimum cell shape when the vector sum of all the forces acting on the model is zero. The result of segmentation is used to confirm the cell membership of the FISH signals and to reject all the signals that lie outside the cell nuclei. Three-dimensional region isolation and labeling technique is used to label and count the FISH signals per cell nucleus. A simple study on the effect of different segmentation methods over a quantitative analysis of FISH signals is also presented.     

Largest contour segmentation: a tool for the localization of spots in confocal images

An accurate determination of the 3-D positions of multiple spots in images obtained by confocal microscopy is essential for the investigation of the spatial distribution of specific components or processes in biological specimens. The position of the centroid, as an estimator for the position of a spot, can be calculated on the basis of all voxels that belong to the domain of the spot. For this calculation a domain that defines which voxels belong to the spot must be delimited. To create a boundary for a domain we developed a 3-D image segmentation procedure: the largest contour segmentation (LCS). This procedure is based on an iterative region-growing procedure around each local maximum of intensity. By means of this procedure the position of each spot was determined accurately and automatically. Qualities of the procedure were evaluated by means of simulated test-images as well as 3-D images of real biological specimens.     

Formation and separation of root border cells

Plant roots release a large number of border cells into the rhizosphere, which are believed to play a key role in root development and health. The formation and loss of these cells from the root cap region is a developmentally regulated process that is also controlled by phytohormones and environmental factors. The separation of border cells involves the complete dissociation of individual cells from each other and from root tissue. This process requires the activity of cell wall-degrading enzymes that solubilize the cell wall connections between cells. We present and discuss the solubilization process with an emphasis on pectin-degrading enzymes as well as the recently discovered root border-like cells of Arabidopsis thaliana.     

Segmentation of complex cell clusters in microscopic images: application to bone marrow samples.

BACKGROUND: Morphologic examination of bone marrow and peripheral blood samples continues to be the cornerstone in diagnostic hematology. In recent years, interest in automatic leukocyte classification using image analysis has increased rapidly. Such systems collect a series of images in which each cell must be segmented accurately to be classified correctly. Although segmentation algorithms have been developed for sparse cells in peripheral blood, the problem of segmenting the complex cell clusters characterizing bone marrow images is harder and has not been addressed previously. METHODS: We present a novel algorithm for segmenting clusters of any number of densely packed cells. The algorithm first oversegments the image into cell subparts. These parts are then assembled into complete cells by solving a combinatorial optimization problem in an efficient way. RESULTS: Our experimental results show that the algorithm succeeds in correctly segmenting densely clustered leukocytes in bone marrow images. CONCLUSIONS: The presented algorithm enables image analysis-based analysis of bone marrow samples for the first time and may also be adopted for other digital cytometric applications where separation of complex cell clusters is required.     

Three-dimensional segmentation of nuclei and mitotic chromosomes for the study of cell divisions in live Drosophila embryos

Drosophila embryogenesis is an established model to investigate mechanisms and genes related to cell divisions in an intact multicellular organism. Progression through the cell cycle phases can be monitored in vivo using fluorescently labeled fusion proteins and time-lapse microscopy. To measure cellular properties in microscopic images, accurate and fast image segmentation methods are a critical prerequisite. To quantify static and dynamic features of interphase nuclei and mitotic chromosomes, we developed a three-dimensional (3D) segmentation method based on multiple level sets. We tested our method on 3D time-series images of live embryos expressing histone-2Av-green fluorescence protein. Our method is robust to low signal-to-noise ratios inherent to high-speed imaging, fluorescent signals in the cytoplasm, and dynamic changes of shape and texture. Comparisons with manual ground-truth segmentations showed that our method achieves more than 90% accuracy on the object as well as voxel levels and performs consistently throughout all cell cycle phases and developmental stages from syncytial blastoderm to postblastoderm mitotic domains.     

Multi-Input Distributed Classifiers for Synthetic Genetic Circuits

For practical construction of complex synthetic genetic networks able to perform elaborate functions it is important to have a pool of relatively simple modules with different functionality which can be compounded together. To complement engineering of very different existing synthetic genetic devices such as switches, oscillators or logical gates, we propose and develop here a design of synthetic multi-input classifier based on a recently introduced distributed classifier concept. A heterogeneous population of cells acts as a single classifier, whose output is obtained by summarizing the outputs of individual cells. The learning ability is achieved by pruning the population, instead of tuning parameters of an individual cell. The present paper is focused on evaluating two possible schemes of multi-input gene classifier circuits. We demonstrate their suitability for implementing a multi-input distributed classifier capable of separating data which are inseparable for single-input classifiers, and characterize performance of the classifiers by analytical and numerical results. The simpler scheme implements a linear classifier in a single cell and is targeted at separable classification problems with simple class borders. A hard learning strategy is used to train a distributed classifier by removing from the population any cell answering incorrectly to at least one training example. The other scheme implements a circuit with a bell-shaped response in a single cell to allow potentially arbitrary shape of the classification border in the input space of a distributed classifier. Inseparable classification problems are addressed using soft learning strategy, characterized by probabilistic decision to keep or discard a cell at each training iteration. We expect that our classifier design contributes to the development of robust and predictable synthetic biosensors, which have the potential to affect applications in a lot of fields, including that of medicine and industry.     

Segmentation of cell nuclei in tissue section analysis

Image segmentation is a critical step in digital picture analysis, especially for that of tissue sections. As the morphology of the cell nuclei provides important biological information, their segmentation is of particular interest. The known segmentation methods are not adequate for segmenting cell nuclei of tissue sections; the reason for this lies in the optical properties of their images. We have developed new gradient methods of segmentation of previously presegmented images by taking these properties into account and by using the approximately circular shape of the cell nuclei as a priori information. In our first technique, the segment method, the images of the nuclei are divided into eight segments, special gradient filters being defined for each segment. This has enabled us to improve the gradient image. After searching for local maxima, the contours of nuclei can be found. In the second method, the method of transformation into the polar coordinate system (PCS), the a priori information serves to define a circular direction field for gradient computation and contour finding. In contrast with the first method, which offers a rapid, general idea about the nuclear shape, the PCS method permits precise segmentation and morphological analysis of the cell nuclei.     

Immunocytochemical localization of cathepsin D in the rat osteoclast.

We performed immunocytochemical localization of cathepsin D in osteoclasts of the proximal growth plate of the rat femurs using both the avidin-biotin-peroxidase complex method for cryo-semi-thin (1 micron) sections and the colloidal gold-labeled IgG method for K4M ultra-thin sections. At the light microscopic level, cathepsin D immunoreactivity in the osteoclasts appeared at the vesicles, granules, and/or small vacuoles. They were distributed throughout the cytoplasm of each cell and were relatively numerous close to the bone surface. This antigen could not be detected at the eroded bone surface. As for other cells, immunoreactivity was seen only in the lysosomes of osteoblast-like cells. Immunoreactivity in the osteoclasts was stronger and greater in the density and number than in osteoblast-like cells. At the electron microscopic level, osteoclasts with well-developed ruffled border possessed numerous cathepsin D-containing lysosomes, vacuoles, and coated vesicle-like structures. Cathepsin D-containing lysosomes fused with cathepsin-negative vacuoles and formed large secondary lysosomes. Osteoclasts with poorly developed ruffled border possessed fewer cathepsin D-containing lysosomes than those with well-developed ruffled border. No immunogold particles were seen in vacuole-like channel expansions of the ruffled borders, between the channels of the ruffled borders, or on the eroded bone surface. These findings demonstrate that osteoclasts contain a large amount of cathepsin D. They suggest that cathepsin D is necessary for osteoclastic bone resorption, that it plays an indirect rather than direct role.