Context based mixture model for cell phase identification in automated fluorescence microscopy

Background  

Automated identification of cell cycle phases of individual live cells in a large population captured via automated fluorescence microscopy technique is important for cancer drug discovery and cell cycle studies. Time-lapse fluorescence microscopy images provide an important method to study the cell cycle process under different conditions of perturbation. Existing methods are limited in dealing with such time-lapse data sets while manual analysis is not feasible. This paper presents statistical data analysis and statistical pattern recognition to perform this task.     

Focus: A Multifaceted Battle Against Cancer: Extracting Knowledge from Chemical Imaging Data Using Computational Algorithms for Digital Cancer Diagnosis

Fourier transform infrared (FTIR) spectroscopic imaging is an emerging microscopy modality for clinical histopathologic diagnoses as well as for biomedical research. Spectral data recorded in this modality are indicative of the underlying, spatially resolved biochemical composition but need computerized algorithms to digitally recognize and transform this information to a diagnostic tool to identify cancer or other physiologic conditions. Statistical pattern recognition forms the backbone of these recognition protocols and can be used for highly accurate results. Aided by biochemical correlations with normal and diseased states and the power of modern computer-aided pattern recognition, this approach is capable of combating many standing questions of traditional histology-based diagnosis models. For example, a simple diagnostic test can be developed to determine cell types in tissue. As a more advanced application, IR spectral data can be integrated with patient information to predict risk of cancer, providing a potential road to precision medicine and personalized care in cancer treatment. The IR imaging approach can be implemented to complement conventional diagnoses, as the samples remain unperturbed and are not destroyed. Despite high potential and utility of this approach, clinical implementation has not yet been achieved due to practical hurdles like speed of data acquisition and lack of optimized computational procedures for extracting clinically actionable information rapidly. The latter problem has been addressed by developing highly efficient ways to process IR imaging data but remains one that has considerable scope for progress. Here, we summarize the major issues and provide practical considerations in implementing a modified Bayesian classification protocol for digital molecular pathology. We hope to familiarize readers with analysis methods in IR imaging data and enable researchers to develop methods that can lead to the use of this promising technique for digital diagnosis of cancer.     

Pattern recognition software and techniques for biological image analysis

The increasing prevalence of automated image acquisition systems is enabling new types of microscopy experiments that generate large image datasets. However, there is a perceived lack of robust image analysis systems required to process these diverse datasets. Most automated image analysis systems are tailored for specific types of microscopy, contrast methods, probes, and even cell types. This imposes significant constraints on experimental design, limiting their application to the narrow set of imaging methods for which they were designed. One of the approaches to address these limitations is pattern recognition, which was originally developed for remote sensing, and is increasingly being applied to the biology domain. This approach relies on training a computer to recognize patterns in images rather than developing algorithms or tuning parameters for specific image processing tasks. The generality of this approach promises to enable data mining in extensive image repositories, and provide objective and quantitative imaging assays for routine use. Here, we provide a brief overview of the technologies behind pattern recognition and its use in computer vision for biological and biomedical imaging. We list available software tools that can be used by biologists and suggest practical experimental considerations to make the best use of pattern recognition techniques for imaging assays.     

Metabolic profiling, metabolomic and metabonomic procedures for NMR spectroscopy of urine, plasma, serum and tissue extracts

Metabolic profiling, metabolomic and metabonomic studies mainly involve the multicomponent analysis of biological fluids, tissue and cell extracts using NMR spectroscopy and/or mass spectrometry (MS). We summarize the main NMR spectroscopic applications in modern metabolic research, and provide detailed protocols for biofluid (urine, serum/plasma) and tissue sample collection and preparation, including the extraction of polar and lipophilic metabolites from tissues. 1H NMR spectroscopic techniques such as standard 1D spectroscopy, relaxation-edited, diffusion-edited and 2D J-resolved pulse sequences are widely used at the analysis stage to monitor different groups of metabolites and are described here. They are often followed by more detailed statistical analysis or additional 2D NMR analysis for biomarker discovery. The standard acquisition time per sample is 4-5 min for a simple 1D spectrum, and both preparation and analysis can be automated to allow application to high-throughput screening for clinical diagnostic and toxicological studies, as well as molecular phenotyping and functional genomics.     

Efficient motif search in ranked lists and applications to variable gap motifs

Sequence elements, at all levels-DNA, RNA and protein, play a central role in mediating molecular recognition and thereby molecular regulation and signaling. Studies that focus on -measuring and investigating sequence-based recognition make use of statistical and computational tools, including approaches to searching sequence motifs. State-of-the-art motif searching tools are limited in their coverage and ability to address large motif spaces. We develop and present statistical and algorithmic approaches that take as input ranked lists of sequences and return significant motifs. The efficiency of our approach, based on suffix trees, allows searches over motif spaces that are not covered by existing tools. This includes searching variable gap motifs-two half sites with a flexible length gap in between-and searching long motifs over large alphabets. We used our approach to analyze several high-throughput measurement data sets and report some validation results as well as novel suggested motifs and motif refinements. We suggest a refinement of the known estrogen receptor 1 motif in humans, where we observe gaps other than three nucleotides that also serve as significant recognition sites, as well as a variable length motif related to potential tyrosine phosphorylation.     

Receptor protein kinases--pattern recognition receptors in plant immunity

Plant innate immunity is activated either upon perception of pathogen-associated molecular patterns (PAMPs) by pattern recognition receptors (PRRs) or upon resistance (R) protein-mediated recognition of pathogen race-specific effector molecules. Although many plant R proteins have been identified, there is only limited knowledge about plant PRRs. Recently, Cyril Zipfel et al. identified a second Arabidopsis leucine-rich repeat receptor protein kinase implicated in PAMP perception, which suggests that several members of this large protein family function as pattern recognition receptors.     

Poly(ethylene glycol acrylate)‐functionalized hydrogels for heparan sulfate oligosaccharide recognition

Heparan sulfates are complex polysaccharides belonging to the family of glycosaminoglycans that participate to the regulation of cell behavior and tissue homeostasis. The biological activities conferred to heparan sulfates are largely dependent on the content and positioning of the sulfate groups along their saccharidic units. At present, identification of particular sulfation patterns in biologically relevant heparan sulfate sequences remains challenging. Although several approaches for structure analysis exist, the complexity of heparan sulfates makes new and original approaches still required. Here, we used molecular imprinting technologies to prepare a library of polyethylene glycol acrylate functionalized hydrogels with the aim to investigate their applicability as specific recognizing systems for fondaparinux, a synthetic pentasaccharide analog to the antithrombin binding site of heparin. Adequate choice of the hydrogel composition and controlling rebinding conditions were important determinants for improving the sulfated oligosaccharide recognition specificity and selectivity. Our results suggest that molecular imprinting approaches could be a possibility for the specific recognition of biologically active sequences in heparan sulfates.     

Automatic protein structure prediction system enabling rapid and accurate model building for enzyme screening

Protein structure prediction has great potential of understanding the function of proteins at the molecular level and designing novel protein functions. Here, we report rapid and accurate structure prediction system running in an automated manner. Since fold recognition of the target protein to be modeled is the starting point of the template-guided model building process, various approaches – such as profile analysis, threading, and SCOP fold classification – have been applied to generate the template library and to select the best template structure. After the best template was determined, fold consistency within the template candidates was considered using TM-score and SCOP database to select additional template structures among the template library. To generate a total of 100 decoy sets, MODELLER was used with the selected template structure. The predicted decoys were clustered with the RMSD deviation criterion of 3 Å to obtain centroids from each cluster. Finally, the selected centroids were subject to side-chain rearrangement using SCWRL module. Our fully automated structure prediction system was examined with sample test sets consisting of recently released 80 PDB chains. Judged by the TM-score (≥0.4), we concluded that 60 cases (75%) showed similar structures of statistical significance. This prediction system provides the users with simple and reliable models within hours of query submission, so that it is quite simply used for high throughput enzyme screening.     

Quantifying Histological Features of Cancer Biospecimens for Biobanking Quality Assurance Using Automated Morphometric Pattern Recognition Image Analysis Algorithms

Biorepository-supported translational research depends on high-quality, well-annotated specimens. Histopathology assessment contributes insight into how representative lesions are for research objectives. Feasibility of documenting histological proportions of tumor and stroma was studied in an effort to enhance information regarding biorepository tissue heterogeneity. Using commercially available software, unique spatial-spectral algorithms were developed for applying automated pattern recognition morphometric image analysis to quantify histologic tumor and nontumor tissue areas in biospecimen tissue sections. Measurements were acquired successfully for 75/75 (100%) lymphomas, 76/77 (98.7%) osteosarcomas, and 60/70 (85.7%) melanomas. The percentage of tissue area occupied by tumor varied among patients and tumor types and was distributed around medians of 94% [interquartile range (IQR)=14%] for lymphomas, 84% for melanomas (IQR=24%), and 39% for osteosarcomas (IQR=44%). Within-patient comparisons from a subset, including multiple individual patient specimens, revealed ≤12% median coefficient of variation (CV) for lymphomas and melanomas. Phenotypic heterogeneity of osteosarcomas resulted in 33% median CV. Uniformly applied, tumor-specific pattern recognition software permits automated tissue-feature quantification. Furthermore, dispersion analyses of area measurements across collections, as well as of multiple specimens from individual patients, support using limited tissue slices to gauge features for some tumor types. Quantitative image analysis automation is anticipated to minimize variability associated with routine biorepository pathologic evaluations and enhance biomarker discovery by helping to guide the selection of study-appropriate specimens.     

Computational Chemical Imaging for Cardiovascular Pathology: Chemical Microscopic Imaging Accurately Determines Cardiac Transplant Rejection

Rejection is a common problem after cardiac transplants leading to significant number of adverse events and deaths, particularly in the first year of transplantation. The gold standard to identify rejection is endomyocardial biopsy. This technique is complex, cumbersome and requires a lot of expertise in the correct interpretation of stained biopsy sections. Traditional histopathology cannot be used actively or quickly during cardiac interventions or surgery. Our objective was to develop a stain-less approach using an emerging technology, Fourier transform infrared (FT-IR) spectroscopic imaging to identify different components of cardiac tissue by their chemical and molecular basis aided by computer recognition, rather than by visual examination using optical microscopy. We studied this technique in assessment of cardiac transplant rejection to evaluate efficacy in an example of complex cardiovascular pathology. We recorded data from human cardiac transplant patients’ biopsies, used a Bayesian classification protocol and developed a visualization scheme to observe chemical differences without the need of stains or human supervision. Using receiver operating characteristic curves, we observed probabilities of detection greater than 95% for four out of five histological classes at 10% probability of false alarm at the cellular level while correctly identifying samples with the hallmarks of the immune response in all cases. The efficacy of manual examination can be significantly increased by observing the inherent biochemical changes in tissues, which enables us to achieve greater diagnostic confidence in an automated, label-free manner. We developed a computational pathology system that gives high contrast images and seems superior to traditional staining procedures. This study is a prelude to the development of real time in situ imaging systems, which can assist interventionists and surgeons actively during procedures.     

Integrative approaches to morphogenesis: lessons from dorsal closure

Although developmental biology has been dominated by the genetic analysis of embryonic development, in recent years genetic tools have been combined with new approaches such as imaging of live processes, automated and quantitative image analysis, mechanical perturbation and mathematical modeling, to study the principles underlying the formation of organisms. Here we focus on recent work carried out on Dorsal Closure, a morphogenetic process during Drosophila embryogenesis, to illustrate how this multidisciplinary approach is yielding new and unexpected insights into how cells organize themselves through the activity of their molecular components to give rise to the stereotyped and macroscopic movements observed during development.     

Soft tissue tumors of the prostate: a review

Prostate cancer is a leading cause of cancer-related death in adult men. Some prostates that are suspected to be involved by prostatic adenocarcinoma or nodular prostatic hyperplasia through clinical examination and imaging studies proves on histologic examination to be a soft tissue tumor. This paper outlines the most common soft tissue tumors of the prostate and categorizes them into benign, malignant or miscellaneous. Pathologists must be aware that most, if not all, soft tissue tumors of the body may also be found in the prostate. Diagnostic immunohistochemistry is an important adjunct to histopathology for proper diagnosis and tumor subclassification.     

Diagnosing benign and malignant lesions in breast tissue sections by using IR-microspectroscopy

The collection of IR spectra through microscope optics and the visualization of the IR data by IR imaging represent a visualization approach, which uses infrared spectral features as a native intrinsic contrast mechanism. To illustrate the potential of this spectroscopic methodology in breast cancer research, we have acquired IR-microspectroscopic data from benign and malignant lesions in breast tissue sections by point microscopy with spot sizes of 30-40 microm. Four classes of distinct breast tissue spectra were defined and stored in the data base: fibroadenoma (a total of 1175 spectra from 14 patients), ductal carcinoma in situ (a total of 1349 spectra from 8 patients), connective tissue (a total of 464 spectra), and adipose tissue (a total of 146 spectra). Artifical neural network analysis, a supervised pattern recognition method, was used to develop an automated classifier to separate the four classes. After training the artifical neural network classifier, infrared spectra of independent external validation data sets ("unknown spectra") were analyzed. In this way, all spectra (a total of 386) taken from micro areas inside the epithelium of fibroadenomas from 4 patients were correctly classified. Out of the 421 spectra taken from micro areas of the in situ component of invasive ductal carcinomas of 3 patients, 93% were correctly identified. Based on these results, the potential of the IR-microspectroscopic approach for diagnosing breast tissue lesions is discussed.     

Gamete recognition in higher plants: an abstruse but charming mystery

Although much effort has been made to uncover the mechanism underlying double fertilization, little knowledge has been acquired for understanding the molecular base of gamete recognition, mainly because of technical limitations. Still,progress has been made in terms of the mechanism, including the identification of candidate molecules that are involved in gamete recognition in angiosperms. New cues for gamete recognition have been found by the successful separation of the gametes and construction of gamete-specific cDNA libraries in several species, and the application of molecular approaches for studying this process by mutations. Thus, the topic is considered an abstruse but charming mystery.     

Quantitative Ultrasound Spectroscopic Imaging for Characterization of Disease Extent in Prostate Cancer Patients

Three-dimensional quantitative ultrasound spectroscopic imaging of prostate was investigated clinically for the noninvasive detection and extent characterization of disease in cancer patients and compared to whole-mount, whole-gland histopathology of radical prostatectomy specimens. Fifteen patients with prostate cancer underwent a volumetric transrectal ultrasound scan before radical prostatectomy. Conventional-frequency (~ 5 MHz) ultrasound images and radiofrequency data were collected from patients. Normalized power spectra were used as the basis of quantitative ultrasound spectroscopy. Specifically, color-coded parametric maps of 0-MHz intercept, midband fit, and spectral slope were computed and used to characterize prostate tissue in ultrasound images. Areas of cancer were identified in whole-mount histopathology specimens, and disease extent was correlated to that estimated from quantitative ultrasound parametric images. Midband fit and 0-MHz intercept parameters were found to be best associated with the presence of disease as located on histopathology whole-mount sections. Obtained results indicated a correlation between disease extent estimated noninvasively based on midband fit parametric images and that identified histopathologically on prostatectomy specimens, with an r² value of 0.71 (P < .0001). The 0-MHz intercept parameter demonstrated a lower level of correlation with histopathology. Spectral slope parametric maps offered no discrimination of disease. Multiple regression analysis produced a hybrid disease characterization model (r² = 0.764, P < .05), implying that the midband fit biomarker had the greatest correlation with the histopathologic extent of disease. This work demonstrates that quantitative ultrasound spectroscopic imaging can be used for detecting prostate cancer and characterizing disease extent noninvasively, with corresponding gross three-dimensional histopathologic correlation.     

Diagnosing benign and malignant lesions in breast tissue sections by using IR-microspectroscopy

The collection of IR spectra through microscope optics and the visualization of the IR data by IR imaging represent a visualization approach, which uses infrared spectral features as a native intrinsic contrast mechanism. To illustrate the potential of this spectroscopic methodology in breast cancer research, we have acquired IR-microspectroscopic data from benign and malignant lesions in breast tissue sections by point microscopy with spot sizes of 30-40 μm. Four classes of distinct breast tissue spectra were defined and stored in the data base: fibroadenoma (a total of 1175 spectra from 14 patients), ductal carcinoma in situ (a total of 1349 spectra from 8 patients), connective tissue (a total of 464 spectra), and adipose tissue (a total of 146 spectra). Artifical neural network analysis, a supervised pattern recognition method, was used to develop an automated classifier to separate the four classes. After training the artifical neural network classifier, infrared spectra of independent external validation data sets (“unknown spectra”) were analyzed. In this way, all spectra (a total of 386) taken from micro areas inside the epithelium of fibroadenomas from 4 patients were correctly classified. Out of the 421 spectra taken from micro areas of the in situ component of invasive ductal carcinomas of 3 patients, 93% were correctly identified. Based on these results, the potential of the IR-microspectroscopic approach for diagnosing breast tissue lesions is discussed.     

Feasibility study on automated recognition of allergenic pollen: grass, birch and mugwort

Quantification of airborne pollen is an important tool in scientific research and patient care in allergy. The currently available method relies on microscopic examination of pollen slides, performed by qualified researchers. Although highly reliable, the method is labor intensive and requires extensive training of the researchers involved. In an approach to develop alternative detection methods, we performed a feasibility study on the automated recognition of the allergenic relevant pollen, grass, birch, and mugwort, by utilizing digital image analysis and pattern recognition tools. Of a total of 254 pollen samples (including 79 of grass, 79 of birch and 96 of mugwort), 97.2% were recognized correctly. This encouraging result provides a promising prospect for future developments.     

Pattern Detection in Intensive Care Monitoring Time Series with Autoregressive Models: Influence of the Model Order

The detection of patterns in monitoring data of vital signs is of great importance for adequate bedside decision support in critical care. Currently used alarm systems, which are based on fixed thresholds and independency assumptions, are not satisfactory in clinical practice. Time series techniques such as AR‐models consider autocorrelations within the series, which can be used for pattern recognition in the data. For practical applications in intensive care the data analysis has to be automated. An important issue is the suitable choice of the model order which is difficult to accomplish online. In a comparative case‐study we analyzed 34564 univariate time series of hemodynamic variables in critically ill patients by autoregressive models of different orders and compared the results of pattern detection. AR(2)‐models seem to be most suitable for the detection of clinically relevant patterns, thus affirming that treating the data as independent leads to false alarms. Moreover, using AR(2)‐models affords only short estimation periods. These findings for pattern detection in intensive care data are of medical importance as they justify a preselection of a model order, easing further automated statistical online analysis.     

Automated identification of Fos expression

The concentration of Fos, a protein encoded by the immediate-early gene c-fos, provides a measure of synaptic activity that may not parallel the electrical activity of neurons. Such a measure is important for the difficult problem of identifying dynamic properties of neuronal circuitries activated by a variety of stimuli and behaviours. We employ two-stage statistical pattern recognition to identify cellular nuclei that express Fos in two-dimensional sections of rat forebrain after administration of antipsychotic drugs. In stage one, we distinguish dark-stained candidate nuclei from image background by a thresholding algorithm and record size and shape measurements of these objects. In stage two, we compare performance of linear and quadratic discriminants, nearest-neighbour and artificial neural network classifiers that employ functions of these measurements to label candidate objects as either Fos nuclei, two touching Fos nuclei or irrelevant background material. New images of neighbouring brain tissue serve as test sets to assess generalizability of the best derived classification rule, as determined by lowest cross-validation misclassification rate. Three experts, two internal and one external, compare manual and automated results for accuracy assessment. Analyses of a subset of images on two separate occasions provide quantitative measures of inter- and intra-expert consistency. We conclude that our automated procedure yields results that compare favourably with those of the experts and thus has potential to remove much of the tedium, subjectivity and irreproducibility of current Fos identification methods in digital microscopy.