Chromosome Image Analysis Research Articles

ChromEDA: Chromosome classification by ensemble framework based domain adaptation.

The precise identification of chromosomal abnormalities based on microscopy images in clinics heavily relies on manual analysis by karyotype experts, which is expensive and labor-intensive. Therefore, there is an urgent need for automatic classification technologies. However, current deep learning-based methods require a large amount of annotated data to achieve satisfactory performance. Additionally, there are domain shifts between datasets obtained from different instruments and clinical agencies, which hinders the utilization of existing public data resources. To overcome these challenges, we propose a novel ensemble learning framework called ChromEDA. This method incorporates soft pseudo-label learning, adversarial learning, and angle classification learning strategies to mitigate the domain shift. To evaluate its performance, we conducted cross-domain classification experiments using both public and private datasets. The results demonstrate that ChromEDA effectively addresses the domain shift issue and outperforms existing methods in cross-domain chromosome classification tasks. RESEARCH HIGHLIGHTS: The research transferred an unsupervised cross-domain algorithm into the chromosome microscopic image processing area, which can support fast and efficient deployment in equipment such as microscopies. To our knowledge, it is the first domain adaption application in chromosome image analysis. The research designed an ensemble learning framework-based chromosome classification model to alleviate the domain shift between different domains. It is a combination of soft pseudo-label learning, adversarial learning, and self-supervised angle recognition, which can significantly reduce the annotation requirements for chromosomal image processing. The results find that for chromosomes sharing the same centromere position category, banding information may contribute more to within-group classification. Thus, banding pattern information is encouraged to be considered for more precise classification of chromosomes.

Automated identification of single and clustered chromosomes for metaphase image analysis

BackgroundChromosome analysis is laborious and time-consuming. Automated methods can significantly increase the efficiency of chromosome analysis. For the automated analysis of chromosome images, single and clustered chromosomes must be identified. Herein, we propose a feature-based method for distinguishing between single chromosomes and clustered chromosome. MethodThe proposed method comprises three main steps. In the first step, chromosome objects are segmented from metaphase chromosome images in advance. In the second step, seven features are extracted from each segmented object, i.e., the normalized area, area/boundary ratio, side branch index, exhaustive thresholding index, normalized minimum width, minimum concave angle, and maximum boundary shift. Finally, the segmented objects are classified as a single chromosome or chromosome cluster using a combination of the seven features. ResultsIn total, 43,391 segmented objects, including 39,892 single chromosomes and 3,499 chromosome clusters, are used to evaluate the proposed method. The results show that the proposed method achieves an accuracy of 98.92% by combining the seven features using support vector machine. ConclusionsThe proposed method is highly effective in distinguishing between single and clustered chromosomes and can be used as a preprocessing procedure for automated chromosome image analysis.

Traditional and deep‐based techniques for end‐to‐end automated karyotyping: A review

AbstractIn the field of cytogenetics, chromosome image analysis or karyotyping from metaphase images plays an imperative role in the diagnosis, prognosis and treatment assessment of different genetic disorders and cancers. This paper is a comprehensive review on different traditional and deep‐based techniques, which are utilized in the design of automated karyotyping systems (AKSs). By this review, a detailed methodology is suggested for the design of end‐to‐end automated karyotyping system (EEAKS) which portrays a sequential multi stage approach. Methods related to all the stages in EEAKS are systematically surveyed by exploring the state of the art literature. Datasets and performance measures incorporated in the past studies are explored. Even though numerous methods were proposed throughout the past three decades, a completely automated framework has not yet been acknowledged. Inferences from this study show that, while various traditional image processing strategies are utilized for pre‐processing and segmentation, machine learning techniques are used only for the classification purpose. In conventional classifiers, artificial neural networks are generally utilized even when the peak performance is given by support vector machines. However, owing to the recent prodigious breakthrough in computer vision, deep neural networks are progressively utilized for developing automated systems. It is seen that deep neural networks are not yet explored in the realm of pre‐processing stage of EEAKS. However, limited number of methods based on convolutional neural networks (CNN) are utilized in all other stages. This review recommends a hybrid CNN for the design of EEAKS, in which all the stages can be automated by sub CNNs. Methodology for generating sufficient datasets is also discussed here which is, indeed, required for further research in this area. This paper concludes with future research directions for the development of a fully automated end‐to‐end karyotyping system.

Imaging approaches for chromosome structures.

This review describes image analyses for chromosome visible structures, focusing on the chromosome imaging system CHIAS (Chromosome Image Analyzing System). CHIAS is the first comprehensive imaging system for the analysis and characterization of plant chromosomes. A simulation method for human vision for capturing band positive regions was developed and used for the image analysis of large plant chromosomes with bands. Applying this method to C-banded Crepis chromosomes enabled recognition of band positive regions as seen by human vision. Furthermore, a new image parameter, condensation pattern was developed and successfully applied to identify small plant chromosomes such as rice and brassicas. Condensation profile (CP) derived from condensation pattern was also effective in developing quantitative chromosome maps. The result was quantitative chromosomal maps of several plants with small chromosomes, including Arabidopsis, diploid brassicas, rapeseed, rice, spinach, and sugarcane. In the final chapter, various applications of imaging techniques to the analysis of pachytene chromosomes, improved visibility of multicolor FISH images, 3D reconstruction of a human chromosome based on cross-section images obtained by a FIB/SEM, automatic extraction of chromosomal regions by machine learning, etc. are described.

Analysis of human chromosome images: Application towards an automated chromosome classification

AbstractKaryotyping is the process of arranging the chromosomes to examine the chromosomal abnormalities, a solution for birth defects. Karyotype is achieved from analyzing the metaspread chromosome images. Metaspread chromosome images generally have problem related with overlapping and touching of chromosomes with each other. Segmentation and separation of overlapping and touching chromosomes are still an open issue. The paper proposes an automated algorithm to separate the touching chromosomes and overlapping chromosomes from G‐Band metaspread images. Segmentation is performed by obtaining the interesting (concave and convex) points on the image contour. A proper hypothesis is performed for separating overlapped and touching chromosome images. The isolation of overlapping and touching metaspread chromosome image helps in accurate automated chromosome classification to obtain a karyotype. The overall accuracy of the proposed method works well for separation of touching chromosome with 97.09% and overlapped chromosome with 95%.

Quantitative analysis of chromosome localization in the nucleus.

The spatial organization of the genome within the interphase nucleus is important for mediating genome functions. The radial organization of chromosome territories has been studied traditionally using two-dimensional fluorescence in situ hybridization (FISH) using labeled whole chromosome probes. Information from 2D-FISH images is analyzed quantitatively and is depicted in the form of the spatial distribution of chromosomes territories. However, to the best of our knowledge no open-access tools are available to delineate the position of chromosome territories from 2D-FISH images. In this chapter we present a methodology termed Image Analysis of Chromosomes for computing their localization (IMACULAT). IMACULAT is an open-access, automated tool that partitions the cell nucleus into shells of equal area or volume and computes the spatial distribution of chromosome territories.

IMACULAT - an open access package for the quantitative analysis of chromosome localization in the nucleus.

The alteration in the location of the chromosomes within the nucleus upon action of internal or external stimuli has been implicated in altering genome function. The effect of stimuli at a whole genome level is studied by using two-dimensional fluorescence in situ hybridization (FISH) to delineate whole chromosome territories within a cell nucleus, followed by a quantitative analysis of the spatial distribution of the chromosome. However, to the best of our knowledge, open access software capable of quantifying spatial distribution of whole chromosomes within cell nucleus is not available. In the current work, we present a software package that computes localization of whole chromosomes - Image Analysis of Chromosomes for computing localization (IMACULAT). We partition the nucleus into concentric elliptical compartments of equal area and the variance in the quantity of any chromosome in these shells is used to determine its localization in the nucleus. The images are pre-processed to remove the smudges outside the cell boundary. Automation allows high throughput analysis for deriving statistics. Proliferating normal human dermal fibroblasts were subjected to standard a two-dimensional FISH to delineate territories for all human chromosomes. Approximately 100 images from each chromosome were analyzed using IMACULAT. The analysis corroborated that these chromosome territories have non-random gene density based organization within the interphase nuclei of human fibroblasts. The ImageMagick Perl API has been used for pre-processing the images. The source code is made available at www.sanchak.com/imaculat.html.

Evaluations of auto-focusing methods under a microscopic imaging modality for metaphase chromosome image analysis.

Background: Auto-focusing is an important operation in high throughput imaging scanning. Although many auto-focusing methods have been developed and tested for a variety of imaging modalities, few investigations have been performed on the selection of an optimal auto-focusing method that is suitable for the pathological metaphase chromosome analysis under a high resolution scanning microscopic system.Objective: The purpose of this study is to investigate and identify an optimal auto-focusing method for the pathological metaphase chromosome analysis.Methods: In this study, five auto-focusing methods were applied and tested using metaphase chromosome images acquired from bone marrow and blood specimens. These methods were assessed by measuring a number of indices including execution time, accuracy, number of false maxima, and full width at half maximum (FWHM).Results: For the specific condition investigated in this study, the results showed that the Brenner gradient and threshold pixel counting methods were the optimal methods for acquiring high quality metaphase chromosome images from the bone marrow and blood specimens, respectively.Conclusions: Selecting an optimal auto-focusing method depends on the specific clinical tasks. This study also provides useful information for the design and implementation of the high throughput microscopic image scanning systems in the future digital pathology.

Integration of Linkage and Chromosome Maps of Red Clover (Trifolium pratense L.)

Red clover (Trifolium pratense L.) is a forage legume and an allogamous diploid plant (2n = 14; 440 Mb). Here, we examine the 7 prometaphase chromosomes of red clover using fluorescence in situ hybridization (FISH) with ribosomal RNA sequences, pericentromeric and telomeric repeats, as well as bacterial artificial chromosome (BAC) clones. Position of hybridization signals and chromosome condensation patterns were quantified by the help of the chromosome image analysis system ver. 4.0 (CHIAS IV). Fourteen BAC clones belonging to linkage groups (LG) 1–7 hybridized to individual chromosomes 4, 2, 6, 5, 1, 7, and 3, respectively. Quantitative analysis of FISH mapping and chromosome analysis using CHIAS IV allowed us to construct a quantitative idiogram that constitutes the comprehensive chromosome map of red clover. Chromosomal positions of the 26S rDNA locus were detected at a heterozygous locus on chromosome 6 in the variety HR, and polymorphisms of rDNA loci were observed in other varieties, although chromosomal positions of some BAC clones did not vary among HR and other varieties. These results demonstrate chromosomal collinearity among allogamous red clover varieties. This integration of genetic linkage and quantitative chromosome maps should provide valuable insight into allogamous legume genetics.

Pilot website to support international collaboration for dose assessments in a radiation emergency

Abstract Nuclear terrorism has emerged as a significant threat which could require timely medical interventions to reduce potential radiation casualties. Early dose assessments are critical since optimal care depends on knowing a victim’s radiation dose. The dicentric chromosome aberration assay is considered the “gold standard” to estimate the radiation dose because the yield of dicentrics correlates positively with the absorbed dose. Dicentrics have a low background frequency, are independent of age and gender and are relatively easy to identify. This diagnostic test for radiation exposure, however, is labor intensive and any single or small group of laboratories could easily be overwhelmed by a mass casualty event. One solution to this potential problem is to link the global WHO BioDoseNet members via the Internet so multiple laboratories could work cooperatively to screen specimens for dicentric chromosomes and generate timely dose estimates. Inter-laboratory comparison studies have shown that analysis of electronic chromosome images viewed on the computer monitor produces scoring accuracy equivalent to viewing live images in the microscope. This functional equivalence was demonstrated during a comparative study involving five laboratories constructing 60 Co gamma ray calibration curves and was further confirmed when comparing results of blind dose estimates submitted by each laboratory. It has been further validated in two recent WHO BioDoseNet trial exercises where 20 metaphase images were shared by e-mail and 50 images were shared on a test website created for this purpose. The Internet-based exercise demonstrated a high level of concordance among 20 expert scorers who evaluated the same 50 metaphase spreads selected to exhibit no, low, moderate and severe radiation damage. Nineteen of 20 scorers produced dicentric equivalent counts within the 95% confidence limits of the mean. The Chi-squared test showed strong evidence of homogeneity in the data ( p = 0.999). Altogether, data obtained from these studies support the conclusion that Internet-based scoring is likely to overcome the “bottleneck” in workflow, reduce turn-a-round time for dose estimates and ultimately strengthen surge capacity. Use of the Internet for biodosimetry would obviously leverage the human and equipment resources throughout the world. As part of radiation emergency planning, we conclude that a global IT network/infrastructure is needed to serve the needs of an expanding biodosimetry community and should be given high priority to meet the growing threat of radiological and nuclear terrorism.

Image analysis of small plant chromosomes by using an improved system, CHIAS IV

Image analysis of small plant chromosomes by using an improved system, CHIAS IV

Automatic Pattern Extraction and Classification for Chromosome Images

Chromosome image analysis and pattern classification is one of the essential tasks in genetic syndrome diagnoses. An automatic procedure is introduced for chromosome image analysis. The pale-path algorithm is proposed to segment touching and overlapping chromosomes. Medial axis is extracted by the middle point algorithm. Chromosome band is enhanced by the algorithm based on multiscale wavelets Bi-spline, and extracted by average gray profile, gradient profile and shape profile. The multilayer classifier is used to classify the chromosome pattern calculated by weighted density distribution algorithm. Experiment results demonstrate that the algorithms perform well.

Advances in rice chromosomes research

Rice chromosome research was initiated around 1910 and has been progressing ever since. Twelve pairs of chromosomes (2n = 24) and a 440 Mb genome, the smallest genome among all main cereal crops, have been identified. According to the recent rice chromosome researches, all chromosomes were well identified over the disadvantages of small chromosome size and quantitative rice chromosomal maps have been developed using chromosome image analysis system (CHIAS). Moreover, agriculturally important genes have been detected using new direct DNA visualization methods with fluorescence in situ hybridization (FISH). This review describes the recent progress in molecular and digital cytogenetics using visualization from a historical perspective, highlighting the future prospects in rice chromosome science. (Communicated by Eishiro SHIKATA, M.J.A.)

Feature analysis and centromere segmentation of human chromosome images using an iterative fuzzy algorithm.

Classification of homologous chromosomes is essential to advanced studies of cancer genetics. Centromere intensities are believed to be an important differentiating feature between homologs. Therefore, segmentation of centromeres is a major step toward the realization of homolog classification. This paper describes an iterative fuzzy algorithm which successfully segments centromeres from images of human chromosomes prepared using fluorescence in-situ hybridization technique. The algorithm is based on assigning a fuzzy membership value to each pixel in the centromere image. An iterative algorithm then updates and minimizes a defined error function. Chromosome 22, a highly heteromorphic chromosome, is used to verify the centromere segmentation method. Homologs of this chromosome are classified based on their segmented centromere intensities as well as their morphological differences. The classification results of these two methods agree completely and are used to validate our developed algorithm.

Karyotype analysis and physical mapping of 18S-5.8S-25S and 5S ribosomal RNA loci in species of genus Lens Miller (Fabaceae)

Abstract The present study on five different species of genus Lens was conducted to prepare the standard karyotypes integrating the 18S-5.8S-25S and 5S rRNA loci using fluorescence in situ hybridisation and computer aided chromosome image analysis. The results revealed that the five species share different karyotypes with minor modifications, possibly as a result of chromosomal interchanges involving exchanges of chromosomal fragments of nearly similar size during the course of evolution of the genus Lens. In each of the five species, a single locus for 18S-5.8S-25S rRNA on a single pair of chromosomes and two loci for 5S rRNA, possibly representing two different repeat units of varying sizes, on two different pairs of chromosomes were mapped. The rRNA loci were localized at different chromosomal positions in different species and could be used as landmarks for chromosome identification. The idiograms based on measurements on chromosomes and the data on physical mapping of rRNA loci allowed clear distinct...

CHIAS-based positioning of recombination hotspots and Giemsa bands in a multireconstructed barley karyotype

Abstract The chromosome complement of reconstructed barley karyotype PK 88 was analyzed by computer-aided Chromosome Image Analysis System (CHIAS). Fine mapping of Giemsa N-bands and regions with increased meiotic recombination activity along each individual chromosome was achieved. It was also found that CHIAS-visualized condensation profiles can be utilized as a reliable criterion for subtle differentiation of hetero- and euchromatin domains within a defined chromosomal regions. Application of CHIAS on karyotypes with distinct chromosome morphology was found to be an appropriate and reliable tool for screening of changes in chromatin compactness and its functional characteristics.

A monophyletic origin of the B chromosomes ofBrachycome dichromosomatica (Asteraceae)

The A and B chromosomes of different karyotype variants (cytodemes A1, A2, A3 and A4) ofBrachycome dichromosomatica were analysed by computer-aided chromosome image analysis and fluorescencein situ hybridisation (FISH). Ribosomal DNA and the B chromosome-specific sequence Bd49 were detected on all B chromosomes. In addition to minor size variation of the Bs, polymorphism of the rDNA and Bd49 position and copy number revealed two major types of B chromosomes. The B chromosomes of all the cytodemes were indistinguishable from each other in length, but that of A3 showed evidence of rearrangements consistent with its long-term geographic isolation. The results presented suggest a monophyletic origin of the B chromosomes ofB. dichromosomatica.

Data-driven homologue matching for chromosome identification.

Karyotyping involves the visualization and classification of chromosomes into standard classes. In "normal" human metaphase spreads, chromosomes occur in homologous pairs for the autosomal classes 1-22, and X chromosome for females. Many existing approaches for performing automated human chromosome image analysis presuppose cell normalcy, containing 46 chromosomes within a metaphase spread with two chromosomes per class. This is an acceptable assumption for routine automated chromosome image analysis. However, many genetic abnormalities are directly linked to structural or numerical aberrations of chromosomes within the metaphase spread. Thus, two chromosomes per class cannot be assumed for anomaly analysis. This paper presents the development of image analysis techniques which are extendible to detecting numerical aberrations evolving from structural abnormalities. Specifically, an approach to identifying "normal" chromosomes from selected class(es) within a metaphase spread is presented. Chromosome assignment to a specific class is initially based on neural networks, followed by banding pattern and centromeric index criteria checking, and concluding with homologue matching. Experimental results are presented comparing neural networks as the sole classifier to our homologue matcher for identifying class 17 within normal and abnormal metaphase spreads.

Toward a completely automatic neural-network-based human chromosome analysis

The application of neural networks (NNs) to automatic analysis of chromosome images is investigated in this paper. All aspects of the analysis, namely segmentation, feature description, selection and extraction, and classification, are studied. As part of the segmentation process, the separation of clusters of partially occluded chromosomes, which is the critical stage that state-of-the-art chromosome analyzers usually fail to accomplish, is performed. First, a moment representation of the image pixels is clustered to create a binary image without a need for threshold selection. Based on the binary image, lines connecting cut points imply possible separations. These hypotheses are verified by a multilayer perceptron (MLP) NN that classifies the two segments created by each separating line. Use of a classification-driven segmentation process gives very promising results without a need for shape modeling or an excessive use of heuristics. In addition, an NN implementation of Sammon's mapping using principal component based initialization is applied to feature extraction, significantly reducing the dimensionality of the feature space and allowing high classification capability. Finally, by applying MLP based hierarchical classification strategies to a well-explored chromosome database, we achieve a classification performance of 83.6%. This is higher than ever published on this database and an improvement of more than 10% in the error rate. Therefore, basing a chromosome analysis on the NN-based techniques that are developed in this research leads toward a completely automatic human chromosome analysis.

Computer-assisted system combining image analysis and chromosome dissection.

The combination of a computerized scanning stage, a UV microbeam, and a chromosome image analysis system allows precise laser dissection of chromosomes and image analysis of the target chromosomes before and after the laser dissection. © 1996 Wiley-Liss, Inc.