Search

Automated image analysis provides a powerful tool when quantifying various characteristics of cell populations. Previously, the validation of image analysis results has been a task of expert biologist, who has manually analyzed the images and provided the ground truth to which the proposed analysis results have been compared. The traditional validation approach, prone to errors and variation, is unfeasible in the emergence of high-throughput measurement systems which make human-based analysis excessively laborious. The systems biology approach for studying, e.g., cellular activity massively in parallel, often lending on high-throughput microscopy, further increases the need for efficient, validated computational methods. As a solution for the problem, we propose a computational framework for simulating fluorescence microscopy images of cell populations. The simulation framework allows generation of synthetic images with realistic characteristics including the ground truth for validation. Thus, the simulation enables validation and performance analysis for various analysis algorithms. By creating a parameterized model of cells based on a given population, the simulator is able to create different cell types. The proposed modular framework, combined with the ability to create high-throughput measurements, provides a powerful tool for validating image analysis methods in traditional microscopy as well as in high content screening. Moreover, we use experimental data to study the validity of the proposed modeling approach.

Fluorescence microscopy combined with digital imaging constructs a basic platform for numerous biomedical studies in the field of cellular imaging. As the studies relying on analysis of digital images have become popular, the validation of image processing methods used in automated image cytometry has become an important topic. Especially, the need for efficient validation has arisen from emerging high-throughput microscopy systems where manual validation is impractical. We present a simulation platform for generating synthetic images of fluorescence-stained cell populations with realistic properties. Moreover, we show that the synthetic images enable the validation of analysis methods for automated image cytometry and comparison of their performance. Finally, we suggest additional usage scenarios for the simulator. The presented simulation framework, with several user-controllable parameters, forms a versatile tool for many kinds of validation tasks, and is freely available at http://www.cs.tut.fi/sgn/csb/simcep.

Synaptic vesicle formation, vesicle activation and exo/endocytosis in the pre-synaptic area are central steps in neuronal communication. The formation and localization of synaptic vesicles in human SH-SY5Y neuroblastoma cells, differentiated with 12-o-tetradecanoyl-phorbol-13-acetate, dibutyryl cyclic AMP, all-trans-retinoic acid (RA) and cholesterol, was studied by fluorescence microscopy and immunocytochemical methods. RA alone or together with cholesterol, produced significant neurite extension and formation of cell-to-cell contacts. Synaptic vesicle formation was followed by anti-synaptophysin (SypI) and AM1-43 staining. SypI was only weakly detected, mainly in cell somata, before 7 days in vitro, after which it was found in neurites. Depolarization of the differentiated cells with high potassium solution increased the number of fluorescent puncta, as well as SypI and AM1-43 co-localization. In addition to increase in the number of synaptic vesicles, RA and cholesterol also increased the number and distribution of lysosome-associated membrane protein 2 labeled lysosomes. RA-induced Golgi apparatus fragmentation was partly avoided by co-treatment with cholesterol. The SH-SY5Y neuroblastoma cell line, differentiated by RA and cholesterol and with good viability in culture, is a valuable tool for basic studies of neuronal metabolism, specifically as a model for dopaminergic neurons.

A new automated image analysis method for quantification of fluorescent dots is presented. This method facilitates counting the number of fluorescent puncta in specific locations of individual cells and also enables estimation of the number of cells by detecting the labeled nuclei. The method is here used for counting the AM1-43 labeled fluorescent puncta in human SH-SY5Y neuroblastoma cells induced to differentiate with all-trans retinoic acid (RA), and further stimulated with high potassium (K+) containing solution. The automated quantification results correlate well with the results obtained manually through visual inspection. The manual method has the disadvantage of being slow, labor-intensive, and subjective, and the results may not be reproducible even in the intra-observer case. The automated method, however, has the advantage of allowing fast quantification with explicitly defined methods, with no user intervention. This ensures objectivity of the quantification. In addition to the number of fluorescent dots, further development of the method allows its use for quantification of several other parameters, such as intensity, size, and shape of the puncta, that are difficult to quantify manually.

Background Several algorithms have been proposed for detecting fluorescently labeled subcellular objects in microscope images. Many of these algorithms have been designed for specific tasks and validated with limited image data. But despite the potential of using extensive comparisons between algorithms to provide useful information to guide method selection and thus more accurate results, relatively few studies have been performed. Results To better understand algorithm performance under different conditions, we have carried out a comparative study including eleven spot detection or segmentation algorithms from various application fields. We used microscope images from well plate experiments with a human osteosarcoma cell line and frames from image stacks of yeast cells in different focal planes. These experimentally derived images permit a comparison of method performance in realistic situations where the number of objects varies within image set. We also used simulated microscope images in order to compare the methods and validate them against a ground truth reference result. Our study finds major differences in the performance of different algorithms, in terms of both object counts and segmentation accuracies. Conclusions These results suggest that the selection of detection algorithms for image based screens should be done carefully and take into account different conditions, such as the possibility of acquiring empty images or images with very few spots. Our inclusion of methods that have not been used before in this context broadens the set of available detection methods and compares them against the current state-of-the-art methods for subcellular particle detection.

Background Fluorescence microscopy is the standard tool for detection and analysis of cellular phenomena. This technique, however, has a number of drawbacks such as the limited number of available fluorescent channels in microscopes, overlapping excitation and emission spectra of the stains, and phototoxicity. Methodology We here present and validate a method to automatically detect cell population outlines directly from bright field images. By imaging samples with several focus levels forming a bright field -stack, and by measuring the intensity variations of this stack over the -dimension, we construct a new two dimensional projection image of increased contrast. With additional information for locations of each cell, such as stained nuclei, this bright field projection image can be used instead of whole cell fluorescence to locate borders of individual cells, separating touching cells, and enabling single cell analysis. Using the popular CellProfiler freeware cell image analysis software mainly targeted for fluorescence microscopy, we validate our method by automatically segmenting low contrast and rather complex shaped murine macrophage cells. Significance The proposed approach frees up a fluorescence channel, which can be used for subcellular studies. It also facilitates cell shape measurement in experiments where whole cell fluorescent staining is either not available, or is dependent on a particular experimental condition. We show that whole cell area detection results using our projected bright field images match closely to the standard approach where cell areas are localized using fluorescence, and conclude that the high contrast bright field projection image can directly replace one fluorescent channel in whole cell quantification. Matlab code for calculating the projections can be downloaded from the supplementary site: http://sites.google.com/site/brightfieldorstaining

An automated image analysis method for identifying folds in tissue section images is presented. Tissue folding is a common artifact in histological images. Folding artifacts form when tissue folds over twice or more when placing it on the microscope slide. As analyzing cell nuclei automatically, the existence of these artifacts causes algorithms easily to give false output. Thus, their identification is essential in order to obtain reliable analysis. The proposed multistage algorithm consists of three phases. First, the section image is converted to HSI color-space and the saturation and intensity components are processed in order to enhance the discrimination of the objective pixels. Next, segmentation is performed using k- means clustering and the cluster containing fold pixels is extracted from the others. Finally, unavoidable segmentation errors caused mostly by the nuclei of similar characteristics with folds are corrected based on the size and component values of the faulty segmented objects. The method is tested on different tissue section images and the results are compared with manually obtained ones with promising results.

Detection and three dimensional reconstruction of cell structures from brightfield microscopy video clips using digital image processing algorithms is presented. While the confocal microscopy offers an efficient technique for three dimensional measurements, extensive and repeated measurements are still often better to be performed using permanent staining and brightfield microscopy. By processing of brightfield microscopy videos using automated and efficient digital image processing algorithms, the tedious task of manual analysis can be avoided. Our two-stage algorithm is applied for 1) cell soma detection and 2) identification of the 3D structure of entire neurons. To verify the results, we present 3D reconstructions of the detected cells.

Synaptic vesicle formation, vesicle activation and exo/endocytosis in the pre-synaptic area are central steps in neuronal communication. The formation and localization of synaptic vesicles in human SH-SY5Y neuroblastoma cells, differentiated with 12-o-tetradecanoyl-phorbol-13-acetate, dibutyryl cyclic AMP, all-trans-retinoic acid (RA) and cholesterol, was studied by fluorescence microscopy and immunocytochemical methods. RA alone or together with cholesterol, produced significant neurite extension and formation of cell-to-cell contacts. Synaptic vesicle formation was followed by anti-synaptophysin (SypI) and AM1-43 staining. SypI was only weakly detected, mainly in cell somata, before 7 days in vitro, after which it was found in neurites. Depolarization of the differentiated cells with high potassium solution increased the number of fluorescent puncta, as well as SypI and AM1-43 co-localization. In addition to increase in the number of synaptic vesicles, RA and cholesterol also increased the number and distribution of lysosome-associated membrane protein 2 labeled lysosomes. RA-induced Golgi apparatus fragmentation was partly avoided by co-treatment with cholesterol. The SH-SY5Y neuroblastoma cell line, differentiated by RA and cholesterol and with good viability in culture, is a valuable tool for basic studies of neuronal metabolism, specifically as a model for dopaminergic neurons.

Subcellular organelles are commonly analyzed using 2D fluorescent microscopy. However, 3D reconstruction and analysis of organelle topology in a high-throughput manner promises to result in a better understanding of cellular systems. We developed image analysis methods for automated quantitative analysis of peroxisome shapes. The methods employ 3D image stacks obtained by confocal microscopy. There are three fundamental phases: image preprocessing and segmentation, 3D reconstruction, and automated quantification of peroxisome topology in 3D using shape descriptors. The algorithms are shown to produce results that can be used to classify objects of different topologies, and to enable visual studies of peroxisomes in 3D.

An automated image analysis method for extract- ing the number of peroxisomes in yeast cells is presented. Two images of the cell population are required for the method: a bright field microscope image from which the yeast cells are detected and the respective fluorescent image from which the number of peroxisomes in each cell is found. The segmentation of the cells is based on clustering the local mean-variance space. The watershed transformation is thereafter employed to separate cells that are clustered together. The peroxisomes are detected by thresholding the fluorescent image. The method is tested with several images of a budding yeast Saccharomyces cerevisiae population, and the results are compared with man- ually obtained results. Index Terms— Yeast, peroxisome biogenesis, image analysis, quantification, segmentation, watershed transformation

We describe a control system to automatically distribute antibody-functionalized beads to addressable assay chambers within a PDMS microfluidic device. The system used real-time image acquisition and processing to manage the valve states required to sort beads with unit precision. The image processing component of the control system correctly counted the number of beads in 99.81% of images (2 689 of 2 694), with only four instances of an incorrect number of beads being sorted to an assay chamber, and one instance of inaccurately counted beads being improperly delivered to waste. Post-experimental refinement of the counting script resulted in one counting error in 2,694 images of beads (99.96% accuracy). We analyzed a range of operational variables (flow pressure, bead concentration, etc.) using a statistical model to characterize those that yielded optimal sorting speed and efficiency. The integrated device was able to capture, count, and deliver beads at a rate of approximately four per minute so that bead arrays could be assembled in 32 individually addressable assay chambers for eight analytical measurements in duplicate (512 beads total) within 2.5 hours. This functionality demonstrates the successful integration of a robust control system with precision bead handling that is the enabling technology for future development of a highly multiplexed bead-based analytical device.

High-throughput cell measurement techniques producing images of cell populations have raised a need for accurate automated image analysis methods. Validating the analysis methods used for automated cytometry is an issue yet to be solved. Manual validation, being an exhaustively laborious task, enables comparison but does not provide solution for large scale analysis. By creating a parametric model for cell shape, and simulating images of cell populations including errors and aberrations caused by the measurement system, validation of different image analysis methods is enabled. As a result, studies with large populations, where the number of cells and many other key parameters are user-tunable, can be carried out by using simulated cell population images. The cell image simulator, as well as validation case studies for segmentation and image restoration are presented.

Neurons cultured in vitro provide a particularly promising experimental system for the analysis of properties, such as information coding, transmission, and learning, that are conventionally associated with biological neural networks. In these systems, isolated cells are placed on top of a recording plate (microelectrode array, MEA), where they spontaneously develop a random connectivity structure. Typical cultures consist of several thousands of neurons and the connectivity density varies from very low at the beginning of an experimental trial to high in mature cultures. In the absence of external stimuli, a culture exhibits a typical pattern of spontaneous activity, alternating intervals of slow spiking and bursting with the transition intervals of increasing activity. Spontaneous activity recorded in the cultures of rat cortical cells is described in [1,2] and an explanation of the phenomena is proposed in [3]. The behavior in the presence of external stimuli is also reported in the literature, for example, the adaptation exhibited in the presence of frequent and rare stimuli is assessed experimentally and through a computational model in [4]. The present work is related to the previously reported study [3] in which an image-processing algorithm is used to detect some structural parameters of cell cultures. A typical result from this study is illustrated in Figure 1. The original image of cultured cells on top of recording plate is shown in panel A, one of its segments in B, and the result of the applied algorithm in C. The blue pattern on panel C corresponds to cells. This approach, in general, enables automated estimation of parameters like the number of cells, or the average density of connections between the cells. Here, we propose a computational model based on the study in [3]. The neural network model is composed of leaky integrate-and-fire neurons, connected in a recurrent network as shown in panel D. The network is fed with the quantitative information about the structure of the cell cultures. Such model, although approximate, captures well the essential properties of the topologies observed in cultures. The presented model is used to reproduce and analyze network behavior observed in the absence of external stimuli. The structural parameters are estimated in different phases of development to closely relate them to the observed behavior. The relation between the network topology and behavior is systematically examined throughout this study.