Automated subcellular location determination and high-throughput microscopy

Automated antinuclear immunofluorescence antibody screening: a comparative study of six computer-aided diagnostic systems

BACKGROUND

Indirect immunofluorescence (IIF) plays an important role in immunological assays for detecting and measuring autoantibodies. However, the method is burdened by some unfavorable features: the need for expert morphologists, the subjectivity of interpretation, and a low degree of standardization and automation. Following the recent statement by the American College of Rheumatology that the IIF technique should be considered as the standard screening method for the detection of anti-nuclear antibodies (ANA), the biomedical industry has developed technological solutions which might significantly improve automation of the procedure, not only in the preparation of substrates and slides, but also in microscope reading.

METHODS

We collected 104 ANA-positive sera from patients with a confirmed clinical diagnosis of autoimmune disease and 40 ANA-negative sera from healthy blood donors. One aliquot of each serum, without information about pattern and titer, was sent to six laboratories of our group, where the sera were tested with the IIF manual method provided by each of the six manufacturers of automatic systems. Assignment of result (pos/neg), of pattern and titer was made by consensus at a meeting attended by all members of the research team. Result was assigned if consensus for pos/neg was reached by at least four of six certifiers, while for the pattern and for the titer, the value observed with higher frequency (mode) was adopted. Seventeen ANA-positive sera and six ANA-negative sera were excluded. Therefore, the study with the following automatic instrumentation was conducted on 92 ANA-positive sera and on 34 ANA-negative sera: Aklides, EUROPattern, G-Sight (I-Sight-IFA), Helios, Image Navigator, and Nova View. Analytical imprecision was measured in five aliquots of the same serum, randomly added to the sample series.

RESULTS

Overall sensitivity of the six automated systems was 96.7% and overall specificity was 89.2%. Most false negatives were recorded for cytoplasmic patterns, whereas among nuclear patterns those with a low level of fluorescence (i.e., multiple nuclear dots, midbody, nuclear rim) were sometimes missed. The intensity values of the light signal of various instruments showed a good correlation with the titer obtained by manual reading (Spearman's rho between 0.672 and 0.839; P<0.0001 for all the systems). Imprecision ranged from 1.99% to 25.2% and, for all the systems, it was lower than that obtained by the manual IIF test (39.1%). The accuracy of pattern recognition, which is for now restricted to the most typical patterns (homogeneous, speckled, nucleolar, centromere, multiple nuclear dots and cytoplasmic) was limited, ranging from 52% to 79%.

CONCLUSIONS

This study, which is the first to compare the diagnostic accuracy of six systems for automated ANA-IIF reading on the same series of sera, showed that all systems are able to perform very well the task for which they were created. Indeed, cumulative automatic discrimination between positive and negative samples had 95% accuracy. All the manufacturers are actively continuing the development of new and more sophisticated software for a better definition in automatic recognition of patterns and light signal conversion in end-point titer. In the future, this may avert the need for serum dilution for titration, which will be a great advantage in economic terms and time-saving.

Building quantitative, three-dimensional atlases of gene expression and morphology at cellular resolution

Animals comprise dynamic three-dimensional arrays of cells that express gene products in intricate spatial and temporal patterns that determine cellular differentiation and morphogenesis. A rigorous understanding of these developmental processes requires automated methods that quantitatively record and analyze complex morphologies and their associated patterns of gene expression at cellular resolution. Here we summarize light microscopy-based approaches to establish permanent, quantitative datasets-atlases-that record this information. We focus on experiments that capture data for whole embryos or large areas of tissue in three dimensions, often at multiple time points. We compare and contrast the advantages and limitations of different methods and highlight some of the discoveries made. We emphasize the need for interdisciplinary collaborations and integrated experimental pipelines that link sample preparation, image acquisition, image analysis, database design, visualization, and quantitative analysis.

PSI: a comprehensive and integrative approach for accurate plant subcellular localization prediction

Predicting the subcellular localization of proteins conquers the major drawbacks of high-throughput localization experiments that are costly and time-consuming. However, current subcellular localization predictors are limited in scope and accuracy. In particular, most predictors perform well on certain locations or with certain data sets while poorly on others. Here, we present PSI, a novel high accuracy web server for plant subcellular localization prediction. PSI derives the wisdom of multiple specialized predictors via a joint-approach of group decision making strategy and machine learning methods to give an integrated best result. The overall accuracy obtained (up to 93.4%) was higher than best individual (CELLO) by ∼10.7%. The precision of each predicable subcellular location (more than 80%) far exceeds that of the individual predictors. It can also deal with multi-localization proteins. PSI is expected to be a powerful tool in protein location engineering as well as in plant sciences, while the strategy employed could be applied to other integrative problems. A user-friendly web server, PSI, has been developed for free access at http://bis.zju.edu.cn/psi/.

Unsupervised clustering of subcellular protein expression patterns in high-throughput microscopy images reveals protein complexes and functional relationships between proteins

Protein subcellular localization has been systematically characterized in budding yeast using fluorescently tagged proteins. Based on the fluorescence microscopy images, subcellular localization of many proteins can be classified automatically using supervised machine learning approaches that have been trained to recognize predefined image classes based on statistical features. Here, we present an unsupervised analysis of protein expression patterns in a set of high-resolution, high-throughput microscope images. Our analysis is based on 7 biologically interpretable features which are evaluated on automatically identified cells, and whose cell-stage dependency is captured by a continuous model for cell growth. We show that it is possible to identify most previously identified localization patterns in a cluster analysis based on these features and that similarities between the inferred expression patterns contain more information about protein function than can be explained by a previous manual categorization of subcellular localization. Furthermore, the inferred cell-stage associated to each fluorescence measurement allows us to visualize large groups of proteins entering the bud at specific stages of bud growth. These correspond to proteins localized to organelles, revealing that the organelles must be entering the bud in a stereotypical order. We also identify and organize a smaller group of proteins that show subtle differences in the way they move around the bud during growth. Our results suggest that biologically interpretable features based on explicit models of cell morphology will yield unprecedented power for pattern discovery in high-resolution, high-throughput microscopy images.

The location of a particular protein in the cell is one of the most important pieces of information that cell biologists use to understand its function. Fluorescent tags are a powerful way to determine the location of a protein in living cells. Nearly a decade ago, a collection of yeast strains was introduced, where in each strain a single protein was tagged with green fluorescent protein (GFP). Here, we show that by training a computer to accurately identify the buds of growing yeast cells, and then making simple fluorescence measurements in context of cell shape and cell stage, the computer could automatically discover most of the localization patterns (nucleus, cytoplasm, mitochondria, etc.) without any prior knowledge of what the patterns might be. Because we made the same, simple measurements for each yeast cell, we could compare and visualize the patterns of fluorescence for the entire collection of strains. This allowed us to identify large groups of proteins moving around the cell in a coordinated fashion, and to identify new, complex patterns that had previously been difficult to describe.

Quantitative image analysis approaches for probing Rab GTPase localization and function in mammalian cells

Membrane traffic pathways play an essential role in cells, providing a mechanism for organelles of the endomembrane system to communicate and exchange material between each other. A significant number of infections and diseases are associated with trafficking pathways, and as such gaining a greater understanding of their regulation is essential. Fluorescence-based imaging techniques are widely used to probe the trafficking machinery within cells, and many of these methods have the potential to be applied in a quantitative manner. In the present mini-review, we highlight several recent examples of how image intensity, kinetic measurements, co-localization and texture feature analysis have been used to study the function of one key family of membrane traffic regulators, the Rab GTPases. We give specific emphasis to the importance of the quantitative nature of these recent studies and comment on their potential applicability to a high-throughput format.

Automated protein subcellular localization based on local invariant features

To understand the function of the encoded proteins, we need to be able to know the subcellular location of a protein. The most common method used for determining subcellular location is fluorescence microscopy which allows subcellular localizations to be imaged in high throughput. Image feature calculation has proven invaluable in the automated analysis of cellular images. This article proposes a novel method named LDPs for feature extraction based on invariant of translation and rotation from given images, the nature which is to count the local difference features of images, and the difference features are given by calculating the D-value between the gray value of the central pixel c and the gray values of eight pixels in the neighborhood. The novel method is tested on two image sets, the first set is which fluorescently tagged protein was endogenously expressed in 10 sebcellular locations, and the second set is which protein was transfected in 11 locations. A SVM was trained and tested for each image set and classification accuracies of 96.7 and 92.3 % were obtained on the endogenous and transfected sets respectively.

Active learning framework with iterative clustering for bioimage classification

Advances in imaging systems have yielded a flood of images into the research field. A semi-automated facility can reduce the laborious task of classifying this large number of images. Here we report the development of a novel framework, CARTA (Clustering-Aided Rapid Training Agent), applicable to bioimage classification that facilitates annotation and selection of features. CARTA comprises an active learning algorithm combined with a genetic algorithm and self-organizing map. The framework provides an easy and interactive annotation method and accurate classification. The CARTA framework enables classification of subcellular localization, mitotic phases and discrimination of apoptosis in images of plant and human cells with an accuracy level greater than or equal to annotators. CARTA can be applied to classification of magnetic resonance imaging of cancer cells or multicolour time-course images after surgery. Furthermore, CARTA can support development of customized features for classification, high-throughput phenotyping and application of various classification schemes dependent on the user's purpose.

Semi-automated imaging systems help with the task of classifying large numbers of biological images. This study presents a novel framework—CARTA—with an active learning algorithm combined with a genetic algorithm, whose applications include the classification of magnetic resonance imaging of cancer cells.

Longitudinal imaging and analysis of neurons expressing polyglutamine-expanded proteins

Misfolded proteins have been implicated in most of the major neurodegenerative diseases, and identifying drugs and pathways that protect neurons from the toxicity of misfolded proteins is of paramount importance. We invented a form of automated imaging and analysis called robotic microscopy that is well suited to the study of neurodegeneration. It enables the monitoring of large cohorts of individual neurons over their lifetimes as they undergo neurodegeneration. With automated analysis, multiple endpoints in neurons can be measured, including survival. Statistical approaches, typically reserved for engineering and clinical medicine, can be applied to these data in an unbiased fashion to discover whether factors contribute positively or negatively to neuronal fate and to quantify the importance of their contribution. Ultimately, multivariate dynamic models can be constructed from these data, which can provide a systems-level understanding of the neurodegenerative disease process and guide the rationale for the development of therapies.

New platform technology for comprehensive serological diagnostics of autoimmune diseases

Antibody assessment is an essential part in the serological diagnosis of autoimmune diseases. However, different diagnostic strategies have been proposed for the work up of sera in particular from patients with systemic autoimmune rheumatic disease (SARD). In general, screening for SARD-associated antibodies by indirect immunofluorescence (IIF) is followed by confirmatory testing covering different assay techniques. Due to lacking automation, standardization, modern data management, and human bias in IIF screening, this two-stage approach has recently been challenged by multiplex techniques particularly in laboratories with high workload. However, detection of antinuclear antibodies by IIF is still recommended to be the gold standard method for antibody screening in sera from patients with suspected SARD. To address the limitations of IIF and to meet the demand for cost-efficient autoantibody screening, automated IIF methods employing novel pattern recognition algorithms for image analysis have been introduced recently. In this respect, the AKLIDES technology has been the first commercially available platform for automated interpretation of cell-based IIF testing and provides multiplexing by addressable microbead immunoassays for confirmatory testing. This paper gives an overview of recently published studies demonstrating the advantages of this new technology for SARD serology.

Automated analysis and reannotation of subcellular locations in confocal images from the Human Protein Atlas

The Human Protein Atlas contains immunofluorescence images showing subcellular locations for thousands of proteins. These are currently annotated by visual inspection. In this paper, we describe automated approaches to analyze the images and their use to improve annotation. We began by training classifiers to recognize the annotated patterns. By ranking proteins according to the confidence of the classifier, we generated a list of proteins that were strong candidates for reexamination. In parallel, we applied hierarchical clustering to group proteins and identified proteins whose annotations were inconsistent with the remainder of the proteins in their cluster. These proteins were reexamined by the original annotators, and a significant fraction had their annotations changed. The results demonstrate that automated approaches can provide an important complement to visual annotation.

Automation in indirect immunofluorescence testing: a new step in the evolution of the autoimmunology laboratory

Indirect immunofluorescence (IIF) plays an important role in immunological and immunometric assays for detecting and measuring autoantibodies. This technology was the first multiplex method used to detect cardinal autoantibodies for the diagnosis of autoimmune diseases. Over the last 20 years, research has enabled the progressive identification of cell and tissue autoantigens which are the target of autoantibodies originally detected by IIF. Accordingly, newer immunometric methods, capable of measuring concentrations of specific autoantibodies directed against these autoantigens, allowed for a gradual replacement of the IIF method in the autoimmunology laboratory. Currently, IIF remains the method of choice only in selected fields of autoimmune diagnostics. Following the recent statement by the American College of Rheumatology that the IIF technique should be considered as the standard screening method for the detection of ANA, the biomedical industry has developed technological solutions which significantly improve automation of the procedure, not only in the preparation of substrates and slides, but also in microscope reading. This review summarizes the general and specific features of new available commercial systems (Aklides, Medipan; Nova View, Inova; Zenit G Sight, A. Menarini Diagnostics; Europattern, Euroimmun; Helios, Aesku.Diagnostics; Image Navigator, Immuno Concepts; Cytospot, Autoimmun Diagnostika) for automation of the IIF method. The expected advantages of automated IIF are the reduction in frequency of false negative and false positive results, the reduction of intra- and inter-laboratory variability, the improvement of correlation of staining patterns with corresponding autoantibody reactivities, and higher throughput in the laboratory workflow.

Biological imaging software tools

Few technologies are more widespread in modern biological laboratories than imaging. Recent advances in optical technologies and instrumentation are providing hitherto unimagined capabilities. Almost all these advances have required the development of software to enable the acquisition, management, analysis and visualization of the imaging data. We review each computational step that biologists encounter when dealing with digital images, the inherent challenges and the overall status of available software for bioimage informatics, focusing on open-source options.

A method for improved clustering and classification of microscopy images using quantitative co-localization coefficients

Background

The localization of proteins to specific subcellular structures in eukaryotic cells provides important information with respect to their function. Fluorescence microscopy approaches to determine localization distribution have proved to be an essential tool in the characterization of unknown proteins, and are now particularly pertinent as a result of the wide availability of fluorescently-tagged constructs and antibodies. However, there are currently very few image analysis options able to effectively discriminate proteins with apparently similar distributions in cells, despite this information being important for protein characterization.

Findings

We have developed a novel method for combining two existing image analysis approaches, which results in highly efficient and accurate discrimination of proteins with seemingly similar distributions. We have combined image texture-based analysis with quantitative co-localization coefficients, a method that has traditionally only been used to study the spatial overlap between two populations of molecules. Here we describe and present a novel application for quantitative co-localization, as applied to the study of Rab family small GTP binding proteins localizing to the endomembrane system of cultured cells.

Conclusions

We show how quantitative co-localization can be used alongside texture feature analysis, resulting in improved clustering of microscopy images. The use of co-localization as an additional clustering parameter is non-biased and highly applicable to high-throughput image data sets.

A multi-label predictor for identifying the subcellular locations of singleplex and multiplex eukaryotic proteins

Subcellular locations of proteins are important functional attributes. An effective and efficient subcellular localization predictor is necessary for rapidly and reliably annotating subcellular locations of proteins. Most of existing subcellular localization methods are only used to deal with single-location proteins. Actually, proteins may simultaneously exist at, or move between, two or more different subcellular locations. To better reflect characteristics of multiplex proteins, it is highly desired to develop new methods for dealing with them. In this paper, a new predictor, called Euk-ECC-mPLoc, by introducing a powerful multi-label learning approach which exploits correlations between subcellular locations and hybridizing gene ontology with dipeptide composition information, has been developed that can be used to deal with systems containing both singleplex and multiplex eukaryotic proteins. It can be utilized to identify eukaryotic proteins among the following 22 locations: (1) acrosome, (2) cell membrane, (3) cell wall, (4) centrosome, (5) chloroplast, (6) cyanelle, (7) cytoplasm, (8) cytoskeleton, (9) endoplasmic reticulum, (10) endosome, (11) extracellular, (12) Golgi apparatus, (13) hydrogenosome, (14) lysosome, (15) melanosome, (16) microsome, (17) mitochondrion, (18) nucleus, (19) peroxisome, (20) spindle pole body, (21) synapse, and (22) vacuole. Experimental results on a stringent benchmark dataset of eukaryotic proteins by jackknife cross validation test show that the average success rate and overall success rate obtained by Euk-ECC-mPLoc were 69.70% and 81.54%, respectively, indicating that our approach is quite promising. Particularly, the success rates achieved by Euk-ECC-mPLoc for small subsets were remarkably improved, indicating that it holds a high potential for simulating the development of the area. As a user-friendly web-server, Euk-ECC-mPLoc is freely accessible to the public at the website http://levis.tongji.edu.cn:8080/bioinfo/Euk-ECC-mPLoc/. We believe that Euk-ECC-mPLoc may become a useful high-throughput tool, or at least play a complementary role to the existing predictors in identifying subcellular locations of eukaryotic proteins.

A multi-functional imaging approach to high-content protein interaction screening

Functional imaging can provide a level of quantification that is not possible in what might be termed traditional high-content screening. This is due to the fact that the current state-of-the-art high-content screening systems take the approach of scaling-up single cell assays, and are therefore based on essentially pictorial measures as assay indicators. Such phenotypic analyses have become extremely sophisticated, advancing screening enormously, but this approach can still be somewhat subjective. We describe the development, and validation, of a prototype high-content screening platform that combines steady-state fluorescence anisotropy imaging with fluorescence lifetime imaging (FLIM). This functional approach allows objective, quantitative screening of small molecule libraries in protein-protein interaction assays. We discuss the development of the instrumentation, the process by which information on fluorescence resonance energy transfer (FRET) can be extracted from wide-field, acceptor fluorescence anisotropy imaging and cross-checking of this modality using lifetime imaging by time-correlated single-photon counting. Imaging of cells expressing protein constructs where eGFP and mRFP1 are linked with amino-acid chains of various lengths (7, 19 and 32 amino acids) shows the two methodologies to be highly correlated. We validate our approach using a small-scale inhibitor screen of a Cdc42 FRET biosensor probe expressed in epidermoid cancer cells (A431) in a 96 microwell-plate format. We also show that acceptor fluorescence anisotropy can be used to measure variations in hetero-FRET in protein-protein interactions. We demonstrate this using a screen of inhibitors of internalization of the transmembrane receptor, CXCR4. These assays enable us to demonstrate all the capabilities of the instrument, image processing and analytical techniques that have been developed. Direct correlation between acceptor anisotropy and donor FLIM is observed for FRET assays, providing an opportunity to rapidly screen proteins, interacting on the nano-meter scale, using wide-field imaging.

WegoLoc: accurate prediction of protein subcellular localization using weighted Gene Ontology terms

SUMMARY

We present an accurate and fast web server, WegoLoc for predicting subcellular localization of proteins based on sequence similarity and weighted Gene Ontology (GO) information. A term weighting method in the text categorization process is applied to GO terms for a support vector machine classifier. As a result, WegoLoc surpasses the state-of-the-art methods for previously used test datasets. WegoLoc supports three eukaryotic kingdoms (animals, fungi and plants) and provides human-specific analysis, and covers several sets of cellular locations. In addition, WegoLoc provides (i) multiple possible localizations of input protein(s) as well as their corresponding probability scores, (ii) weights of GO terms representing the contribution of each GO term in the prediction, and (iii) a BLAST E-value for the best hit with GO terms. If the similarity score does not meet a given threshold, an amino acid composition-based prediction is applied as a backup method.

AVAILABILITY

WegoLoc and User's guide are freely available at the website http://www.btool.org/WegoLoc

CONTACT

smchiks@ks.ac.kr; dougnam@unist.ac.kr

SUPPLEMENTARY INFORMATION

Supplementary data is available at http://www.btool.org/WegoLoc.

Improving the prediction of sub-cellular locations of proteins with a particle swarm optimization-based boosting strategy

Learning from imbalanced data sets presents an important challenge to the machine learning community. Traditional classification methods, seeking to minimize the overall error rate of the whole training set, do not perform well on imbalanced data since they assume a relatively balanced class distribution and put too much strength on the majority class. This is a common scenario when predicting sub-cellular locations of proteins since proteins belonging to certain specific locations are naturally more abundant or have been more extensively studied. In this work, a new method to learn from imbalanced data, called SwarmBoost, is proposed in order to reduce overlapping and noise of imbalanced datasets and improve prediction performances. The method combines oversampling, subsampling based on particle swarm optimization and ensemble methods. Our results show that SwarmBoost equals and in several cases outperforms other common boosting algorithms like DataBoost-Im and AdaBoost, constituting a useful tool for improving sub-cellular location predictions.

Drosophila gene expression pattern annotation through multi-instance multi-label learning

In the studies of Drosophila embryogenesis, a large number of two-dimensional digital images of gene expression patterns have been produced to build an atlas of spatio-temporal gene expression dynamics across developmental time. Gene expressions captured in these images have been manually annotated with anatomical and developmental ontology terms using a controlled vocabulary (CV), which are useful in research aimed at understanding gene functions, interactions, and networks. With the rapid accumulation of images, the process of manual annotation has become increasingly cumbersome, and computational methods to automate this task are urgently needed. However, the automated annotation of embryo images is challenging. This is because the annotation terms spatially correspond to local expression patterns of images, yet they are assigned collectively to groups of images and it is unknown which term corresponds to which region of which image in the group. In this paper, we address this problem using a new machine learning framework, Multi-Instance Multi-Label (MIML) learning. We first show that the underlying nature of the annotation task is a typical MIML learning problem. Then, we propose two support vector machine algorithms under the MIML framework for the task. Experimental results on the FlyExpress database (a digital library of standardized Drosophila gene expression pattern images) reveal that the exploitation of MIML framework leads to significant performance improvement over state-of-the-art approaches.

Ensemble of Neural Networks for Automated Cell Phenotype Image Classification

Subcellular location is related to the knowledge of the spatial distribution of a protein within the cell. The knowledge of the location of all proteins is crucial for several applications ranging from early diagnosis of a disease to monitoring of therapeutic effectiveness of drugs. This chapter focuses on the study of machine learning techniques for cell phenotype image classification and is aimed at pointing out some of the advantages of using a multi-classifier system instead of a stand-alone method to solve this difficult classification problem. The main problems and solutions proposed in this field are discussed and a new approach is proposed based on ensemble of neural networks trained by local and global features. Finally, the most used benchmarks for this problem are presented and an experimental comparison among several state-of-the-art approaches is reported which allows to quantify the performance improvement obtained by the approach proposed in this chapter.

New concepts for building vocabulary for cell image ontologies

Background

There are significant challenges associated with the building of ontologies for cell biology experiments including the large numbers of terms and their synonyms. These challenges make it difficult to simultaneously query data from multiple experiments or ontologies. If vocabulary terms were consistently used and reused across and within ontologies, queries would be possible through shared terms. One approach to achieving this is to strictly control the terms used in ontologies in the form of a pre-defined schema, but this approach limits the individual researcher's ability to create new terms when needed to describe new experiments.

Results

Here, we propose the use of a limited number of highly reusable common root terms, and rules for an experimentalist to locally expand terms by adding more specific terms under more general root terms to form specific new vocabulary hierarchies that can be used to build ontologies. We illustrate the application of the method to build vocabularies and a prototype database for cell images that uses a visual data-tree of terms to facilitate sophisticated queries based on a experimental parameters. We demonstrate how the terminology might be extended by adding new vocabulary terms into the hierarchy of terms in an evolving process. In this approach, image data and metadata are handled separately, so we also describe a robust file-naming scheme to unambiguously identify image and other files associated with each metadata value. The prototype database http://sbd.nist.gov/ consists of more than 2000 images of cells and benchmark materials, and 163 metadata terms that describe experimental details, including many details about cell culture and handling. Image files of interest can be retrieved, and their data can be compared, by choosing one or more relevant metadata values as search terms. Metadata values for any dataset can be compared with corresponding values of another dataset through logical operations.

Conclusions

Organizing metadata for cell imaging experiments under a framework of rules that include highly reused root terms will facilitate the addition of new terms into a vocabulary hierarchy and encourage the reuse of terms. These vocabulary hierarchies can be converted into XML schema or RDF graphs for displaying and querying, but this is not necessary for using it to annotate cell images. Vocabulary data trees from multiple experiments or laboratories can be aligned at the root terms to facilitate query development. This approach of developing vocabularies is compatible with the major advances in database technology and could be used for building the Semantic Web.

TimeLapseAnalyzer: multi-target analysis for live-cell imaging and time-lapse microscopy

The direct observation of cells over time using time-lapse microscopy can provide deep insights into many important biological processes. Reliable analyses of motility, proliferation, invasive potential or mortality of cells are essential to many studies involving live cell imaging and can aid in biomarker discovery and diagnostic decisions. Given the vast amount of image- and time-series data produced by modern microscopes, automated analysis is a key feature to capitalize the potential of time-lapse imaging devices. To provide fast and reproducible analyses of multiple aspects of cell behaviour, we developed TimeLapseAnalyzer. Apart from general purpose image enhancements and segmentation procedures, this extensible, self-contained, modular cross-platform package provides dedicated modalities for fast and reliable analysis of multi-target cell tracking, scratch wound healing analysis, cell counting and tube formation analysis in high throughput screening of live-cell experiments. TimeLapseAnalyzer is freely available (MATLAB, Open Source) at http://www.informatik.uni-ulm.de/ni/mitarbeiter/HKestler/tla.

An automated fluorescence videomicroscopy assay for the detection of mitotic catastrophe

Mitotic catastrophe can be defined as a cell death mode that occurs during or shortly after a prolonged/aberrant mitosis, and can show apoptotic or necrotic features. However, conventional procedures for the detection of apoptosis or necrosis, including biochemical bulk assays and cytofluorometric techniques, cannot discriminate among pre-mitotic, mitotic and post-mitotic death, and hence are inappropriate to monitor mitotic catastrophe. To address this issue, we generated isogenic human colon carcinoma cell lines that differ in ploidy and p53 status, yet express similar amounts of fluorescent biosensors that allow for the visualization of chromatin (histone H2B coupled to green fluorescent protein (GFP)) and centrosomes (centrin coupled to the Discosoma striata red fluorescent protein (DsRed)). By combining high-resolution fluorescence videomicroscopy and automated image analysis, we established protocols and settings for the simultaneous assessment of ploidy, mitosis, centrosome number and cell death (which in our model system occurs mainly by apoptosis). Time-lapse videomicroscopy showed that this approach can be used for the high-throughput detection of mitotic catastrophe induced by three mechanistically distinct anti-mitotic agents (dimethylenastron (DIMEN), nocodazole (NDZ) and paclitaxel (PTX)), and – in this context – revealed an important role of p53 in the control of centrosome number.

Phenotypic profiling of the human genome reveals gene products involved in plasma membrane targeting of SRC kinases

SRC proteins are non-receptor tyrosine kinases that play key roles in regulating signal transduction by a diverse set of cell surface receptors. They contain N-terminal SH4 domains that are modified by fatty acylation and are functioning as membrane anchors. Acylated SH4 domains are both necessary and sufficient to mediate specific targeting of SRC kinases to the inner leaflet of plasma membranes. Intracellular transport of SRC kinases to the plasma membrane depends on microdomains into which SRC kinases partition upon palmitoylation. In the present study, we established a live-cell imaging screening system to identify gene products involved in plasma membrane targeting of SRC kinases. Based on siRNA arrays and a human model cell line expressing two kinds of SH4 reporter molecules, we conducted a genome-wide analysis of SH4-dependent protein targeting using an automated microscopy platform. We identified and validated 54 gene products whose down-regulation causes intracellular retention of SH4 reporter molecules. To detect and quantify this phenotype, we developed a software-based image analysis tool. Among the identified gene products, we found factors involved in lipid metabolism, intracellular transport, and cellular signaling processes. Furthermore, we identified proteins that are either associated with SRC kinases or are related to various known functions of SRC kinases such as other kinases and phosphatases potentially involved in SRC-mediated signal transduction. Finally, we identified gene products whose function is less defined or entirely unknown. Our findings provide a major resource for future studies unraveling the molecular mechanisms that underlie proper targeting of SRC kinases to the inner leaflet of plasma membranes.

A multi-label classifier for predicting the subcellular localization of gram-negative bacterial proteins with both single and multiple sites

Prediction of protein subcellular localization is a challenging problem, particularly when the system concerned contains both singleplex and multiplex proteins. In this paper, by introducing the “multi-label scale” and hybridizing the information of gene ontology with the sequential evolution information, a novel predictor called iLoc-Gneg is developed for predicting the subcellular localization of Gram-positive bacterial proteins with both single-location and multiple-location sites. For facilitating comparison, the same stringent benchmark dataset used to estimate the accuracy of Gneg-mPLoc was adopted to demonstrate the power of iLoc-Gneg. The dataset contains 1,392 Gram-negative bacterial proteins classified into the following eight locations: (1) cytoplasm, (2) extracellular, (3) fimbrium, (4) flagellum, (5) inner membrane, (6) nucleoid, (7) outer membrane, and (8) periplasm. Of the 1,392 proteins, 1,328 are each with only one subcellular location and the other 64 are each with two subcellular locations, but none of the proteins included has pairwise sequence identity to any other in a same subset (subcellular location). It was observed that the overall success rate by jackknife test on such a stringent benchmark dataset by iLoc-Gneg was over 91%, which is about 6% higher than that by Gneg-mPLoc. As a user-friendly web-server, iLoc-Gneg is freely accessible to the public at http://icpr.jci.edu.cn/bioinfo/iLoc-Gneg. Meanwhile, a step-by-step guide is provided on how to use the web-server to get the desired results. Furthermore, for the user's convenience, the iLoc-Gneg web-server also has the function to accept the batch job submission, which is not available in the existing version of Gneg-mPLoc web-server. It is anticipated that iLoc-Gneg may become a useful high throughput tool for Molecular Cell Biology, Proteomics, System Biology, and Drug Development.

Image-derived, three-dimensional generative models of cellular organization

Given the importance of subcellular location to protein function, computational simulations of cell behaviors will ultimately require the ability to model the distributions of proteins within organelles and other structures. Toward this end, statistical learning methods have previously been used to build models of sets of two-dimensional microscope images, where each set contains multiple images for a single subcellular location pattern. The model learned from each set of images not only represents the pattern but also captures the variation in that pattern from cell to cell. The models consist of sub-models for nuclear shape, cell shape, organelle size and shape, and organelle distribution relative to nuclear and cell boundaries, and allow synthesis of images with the expectation that they are drawn from the same underlying statistical distribution as the images used to train them. Here we extend this generative models approach to three dimensions using a similar framework, permitting protein subcellular locations to be described more accurately. Models of different patterns can be combined to yield a synthetic multi-channel image containing as many proteins as desired, something that is difficult to obtain by direct microscope imaging for more than a few proteins. In addition, the model parameters represent a more compact and interpretable way of communicating subcellular patterns than descriptive image features and may be particularly effective for automated identification of changes in subcellular organization caused by perturbagens.

iLoc-Euk: a multi-label classifier for predicting the subcellular localization of singleplex and multiplex eukaryotic proteins

Predicting protein subcellular localization is an important and difficult problem, particularly when query proteins may have the multiplex character, i.e., simultaneously exist at, or move between, two or more different subcellular location sites. Most of the existing protein subcellular location predictor can only be used to deal with the single-location or “singleplex” proteins. Actually, multiple-location or “multiplex” proteins should not be ignored because they usually posses some unique biological functions worthy of our special notice. By introducing the “multi-labeled learning” and “accumulation-layer scale”, a new predictor, called iLoc-Euk, has been developed that can be used to deal with the systems containing both singleplex and multiplex proteins. As a demonstration, the jackknife cross-validation was performed with iLoc-Euk on a benchmark dataset of eukaryotic proteins classified into the following 22 location sites: (1) acrosome, (2) cell membrane, (3) cell wall, (4) centriole, (5) chloroplast, (6) cyanelle, (7) cytoplasm, (8) cytoskeleton, (9) endoplasmic reticulum, (10) endosome, (11) extracellular, (12) Golgi apparatus, (13) hydrogenosome, (14) lysosome, (15) melanosome, (16) microsome (17) mitochondrion, (18) nucleus, (19) peroxisome, (20) spindle pole body, (21) synapse, and (22) vacuole, where none of proteins included has pairwise sequence identity to any other in a same subset. The overall success rate thus obtained by iLoc-Euk was 79%, which is significantly higher than that by any of the existing predictors that also have the capacity to deal with such a complicated and stringent system. As a user-friendly web-server, iLoc-Euk is freely accessible to the public at the web-site http://icpr.jci.edu.cn/bioinfo/iLoc-Euk. It is anticipated that iLoc-Euk may become a useful bioinformatics tool for Molecular Cell Biology, Proteomics, System Biology, and Drug Development Also, its novel approach will further stimulate the development of predicting other protein attributes.

Automated microscopy and image analysis for androgen receptor function

Systems-level approaches have emerged that rely on analytical, microscopy-based technology for the discovery of novel drug targets and the mechanisms driving AR signaling, transcriptional activity, and ligand independence. Single cell behavior can be quantified by high-throughput microscopy methods through analysis of endogenous protein levels and localization or creation of biosensor cell lines that can simultaneously detect both acute and latent responses to known and unknown androgenic stimuli. The cell imaging and analytical protocols can be automated to discover agonist/antagonist response windows for nuclear translocation, reporter gene activity, nuclear export, and subnuclear transcription events, facilitating access to a multiplex model system that is inherently unavailable through classic biochemical approaches. In this chapter, we highlight the key steps needed for developing, conducting, and analyzing high-throughput screens to identify effectors of AR signaling.

EXTRACTING BIOMEDICALLY IMPORTANT INFORMATION FROM LARGE, AUTOMATED IMAGING EXPERIMENTS

Major challenges remain in the extraction of rich information from high-throughput microscopy experiments. In this paper, I describe some of these challenges, particularly those that are the subject of ongoing research in my laboratory. The challenges include segmenting neurons, co-cultures of different cell types, and whole organisms; segmenting and tracking cells in time-lapse images; quantifying complex phenotypic changes; and discovering biologically relevant subpopulations of cells.

A hybrid blob-slice model for accurate and efficient detection of fluorescence labeled nuclei in 3D

BACKGROUND

To exploit the flood of data from advances in high throughput imaging of optically sectioned nuclei, image analysis methods need to correctly detect thousands of nuclei, ideally in real time. Variability in nuclear appearance and undersampled volumetric data make this a challenge.

RESULTS

We present a novel 3D nuclear identification method, which subdivides the problem, first segmenting nuclear slices within each 2D image plane, then using a shape model to assemble these slices into 3D nuclei. This hybrid 2D/3D approach allows accurate accounting for nuclear shape but exploits the clear 2D nuclear boundaries that are present in sectional slices to avoid the computational burden of fitting a complex shape model to volume data. When tested over C. elegans, Drosophila, zebrafish and mouse data, our method yielded 0 to 3.7% error, up to six times more accurate as well as being 30 times faster than published performances. We demonstrate our method's potential by reconstructing the morphogenesis of the C. elegans pharynx. This is an important and much studied developmental process that could not previously be followed at this single cell level of detail.

CONCLUSIONS

Because our approach is specialized for the characteristics of optically sectioned nuclear images, it can achieve superior accuracy in significantly less time than other approaches. Both of these characteristics are necessary for practical analysis of overwhelmingly large data sets where processing must be scalable to hundreds of thousands of cells and where the time cost of manual error correction makes it impossible to use data with high error rates. Our approach is fast, accurate, available as open source software and its learned shape model is easy to retrain. As our pharynx development example shows, these characteristics make single cell analysis relatively easy and will enable novel experimental methods utilizing complex data sets.

Systems microscopy approaches to understand cancer cell migration and metastasis

Cell migration is essential in a number of processes, including wound healing, angiogenesis and cancer metastasis. Especially, invasion of cancer cells in the surrounding tissue is a crucial step that requires increased cell motility. Cell migration is a well-orchestrated process that involves the continuous formation and disassembly of matrix adhesions. Those structural anchor points interact with the extra-cellular matrix and also participate in adhesion-dependent signalling. Although these processes are essential for cancer metastasis, little is known about the molecular mechanisms that regulate adhesion dynamics during tumour cell migration. In this review, we provide an overview of recent advanced imaging strategies together with quantitative image analysis that can be implemented to understand the dynamics of matrix adhesions and its molecular components in relation to tumour cell migration. This dynamic cell imaging together with multiparametric image analysis will help in understanding the molecular mechanisms that define cancer cell migration.

Image analysis benchmarking methods for high-content screen design

The recent development of complex chemical and small interfering RNA (siRNA) collections has enabled large-scale cell-based phenotypic screening. High-content and high-throughput imaging are widely used methods to record phenotypic data after chemical and small interfering RNA treatment, and numerous image processing and analysis methods have been used to quantify these phenotypes. Currently, there are no standardized methods for evaluating the effectiveness of new and existing image processing and analysis tools for an arbitrary screening problem. We generated a series of benchmarking images that represent commonly encountered variation in high-throughput screening data and used these image standards to evaluate the robustness of five different image analysis methods to changes in signal-to-noise ratio, focal plane, cell density and phenotype strength. The analysis methods that were most reliable, in the presence of experimental variation, required few cells to accurately distinguish phenotypic changes between control and experimental data sets. We conclude that by applying these simple benchmarking principles an a priori estimate of the image acquisition requirements for phenotypic analysis can be made before initiating an image-based screen. Application of this benchmarking methodology provides a mechanism to significantly reduce data acquisition and analysis burdens and to improve data quality and information content.

Descriptive no more: the dawn of high-throughput microscopy

The next revolution in microscopy is upon us: it is High-Throughput Imaging (HTI). In HTI large numbers of images from many samples are acquired and analyzed. This has become possible due to a confluence of dramatic progress in microscope engineering, enabling efficient image collection, and the availability of high computing power for data analysis. Combining HTI with genome-wide RNA interference (RNAi)-based gene knockdown technology offers a powerful approach for unbiased discovery of cellular mechanisms.

CellCognition: time-resolved phenotype annotation in high-throughput live cell imaging

Fluorescence time-lapse imaging has become a powerful tool to investigate complex dynamic processes such as cell division or intracellular trafficking. Automated microscopes generate time-resolved imaging data at high throughput, yet tools for quantification of large-scale movie data are largely missing. Here we present CellCognition, a computational framework to annotate complex cellular dynamics. We developed a machine-learning method that combines state-of-the-art classification with hidden Markov modeling for annotation of the progression through morphologically distinct biological states. Incorporation of time information into the annotation scheme was essential to suppress classification noise at state transitions and confusion between different functional states with similar morphology. We demonstrate generic applicability in different assays and perturbation conditions, including a candidate-based RNA interference screen for regulators of mitotic exit in human cells. CellCognition is published as open source software, enabling live-cell imaging-based screening with assays that directly score cellular dynamics.