The recent advent of high-throughput microarray data has enabled the global analysis of the transcriptome, driving the development and application of computational approaches to study transcriptional regulation on the genome scale, by reconstructing in silico the regulatory interactions of the gene network. Although there are many in-depth reviews of such ‘reverse-engineering’ methodologies, most have focused on the practical aspect of data mining, and few on the biological problem and the biological relevance of the methodology. Therefore, in this review, from a biological perspective, we used a set of yeast microarray data as a working example, to evaluate the fundamental assumptions implicit in associating transcription factor (TF)–target gene expression levels and estimating TFs’ activity, and further explore cooperative models. Finally we confirm that the detailed transcription mechanism is overly-complex for expression data alone to reveal, nevertheless, future network reconstruction studies could benefit from the incorporation of context-specific information, the modeling of multiple layers of regulation (e.g. micro-RNA), or the development of approaches for context-dependent analysis, to uncover the mechanisms of gene regulation.

Organisms usually cope with change in the environment by altering the dynamic trajectory of gene expression to adjust the complement of active proteins. The identification of particular sets of genes whose expression is adaptive in response to environmental changes helps to understand the mechanistic base of gene–environment interactions essential for organismic development. We describe a computational framework for clustering the dynamics of gene expression in distinct environments through Gaussian mixture fitting to the expression data measured at a set of discrete time points. We outline a number of quantitative testable hypotheses about the patterns of dynamic gene expression in changing environments and gene–environment interactions causing developmental differentiation. The future directions of gene clustering in terms of incorporations of the latest biological discoveries and statistical innovations are discussed. We provide a set of computational tools that are applicable to modeling and analysis of dynamic gene expression data measured in multiple environments.

Network motifs are statistically overrepresented sub-structures (sub-graphs) in a network, and have been recognized as ‘the simple building blocks of complex networks’. Study of biological network motifs may reveal answers to many important biological questions. The main difficulty in detecting larger network motifs in biological networks lies in the facts that the number of possible sub-graphs increases exponentially with the network or motif size (node counts, in general), and that no known polynomial-time algorithm exists in deciding if two graphs are topologically equivalent. This article discusses the biological significance of network motifs, the motivation behind solving the motif-finding problem, and strategies to solve the various aspects of this problem. A simple classification scheme is designed to analyze the strengths and weaknesses of several existing algorithms. Experimental results derived from a few comparative studies in the literature are discussed, with conclusions that lead to future research directions.

Creating useful software is a major activity of many scientists, including bioinformaticians. Nevertheless, software development in an academic setting is often unsystematic, which can lead to problems associated with maintenance and long-term availibility. Unfortunately, well-documented software development methodology is difficult to adopt, and technical measures that directly improve bioinformatic programming have not been described comprehensively. We have examined 22 software projects and have identified a set of practices for software development in an academic environment. We found them useful to plan a project, support the involvement of experts (e.g. experimentalists), and to promote higher quality and maintainability of the resulting programs. This article describes 12 techniques that facilitate a quick start into software engineering. We describe 3 of the 22 projects in detail and give many examples to illustrate the usage of particular techniques. We expect this toolbox to be useful for many bioinformatics programming projects and to the training of scientific programmers.

Recent advances in massively parallel sequencing technology have created new opportunities to probe the hidden world of microbes. Taxonomy-independent clustering of the 16S rRNA gene is usually the first step in analyzing microbial communities. Dozens of algorithms have been developed in the last decade, but a comprehensive benchmark study is lacking. Here, we survey algorithms currently used by microbiologists, and compare seven representative methods in a large-scale benchmark study that addresses several issues of concern. A new experimental protocol was developed that allows different algorithms to be compared using the same platform, and several criteria were introduced to facilitate a quantitative evaluation of the clustering performance of each algorithm. We found that existing methods vary widely in their outputs, and that inappropriate use of distance levels for taxonomic assignments likely resulted in substantial overestimates of biodiversity in many studies. The benchmark study identified our recently developed ESPRIT-Tree, a fast implementation of the average linkage-based hierarchical clustering algorithm, as one of the best algorithms available in terms of computational efficiency and clustering accuracy.

Over the past two decades, there has been a long-standing debate about the impact of taxon sampling on phylogenetic inference. Studies have been based on both real and simulated data sets, within actual and theoretical contexts, and using different inference methods, to study the impact of taxon sampling. In some cases, conflicting conclusions have been drawn for the same data set. The main questions explored in studies to date have been about the effects of using sparse data, adding new taxa, including more characters from genome sequences and using different (or concatenated) locus regions. These questions can be reduced to more fundamental ones about the assessment of data quality and the design guidelines of taxon sampling in phylogenetic inference experiments. This review summarizes progress to date in understanding the impact of taxon sampling on the accuracy of phylogenetic analysis.

Genetic imprinting, by which the expression of a gene depends on the parental origin of its alleles, may be subjected to reprogramming through each generation. Currently, such reprogramming is limited to qualitative description only, lacking more precise quantitative estimation for its extent, pattern and mechanism. Here, we present a computational framework for analyzing the magnitude of genetic imprinting and its transgenerational inheritance mode. This quantitative model is based on the breeding scheme of reciprocal backcrosses between reciprocal F1 hybrids and original inbred parents, in which the transmission of genetic imprinting across generations can be tracked. We define a series of quantitative genetic parameters that describe the extent and transmission mode of genetic imprinting and further estimate and test these parameters within a genetic mapping framework using a new powerful computational algorithm. The model and algorithm described will enable geneticists to identify and map imprinted quantitative trait loci and dictate a comprehensive atlas of developmental and epigenetic mechanisms related to genetic imprinting. We illustrate the new discovery of the role of genetic imprinting in regulating hyperoxic acute lung injury survival time using a mouse reciprocal backcross design.

In analysis of bioinformatics data, a unique challenge arises from the high dimensionality of measurements. Without loss of generality, we use genomic study with gene expression measurements as a representative example but note that analysis techniques discussed in this article are also applicable to other types of bioinformatics studies. Principal component analysis (PCA) is a classic dimension reduction approach. It constructs linear combinations of gene expressions, called principal components (PCs). The PCs are orthogonal to each other, can effectively explain variation of gene expressions, and may have a much lower dimensionality. PCA is computationally simple and can be realized using many existing software packages. This article consists of the following parts. First, we review the standard PCA technique and their applications in bioinformatics data analysis. Second, we describe recent ‘non-standard’ applications of PCA, including accommodating interactions among genes, pathways and network modules and conducting PCA with estimating equations as opposed to gene expressions. Third, we introduce several recently proposed PCA-based techniques, including the supervised PCA, sparse PCA and functional PCA. The supervised PCA and sparse PCA have been shown to have better empirical performance than the standard PCA. The functional PCA can analyze time-course gene expression data. Last, we raise the awareness of several critical but unsolved problems related to PCA. The goal of this article is to make bioinformatics researchers aware of the PCA technique and more importantly its most recent development, so that this simple yet effective dimension reduction technique can be better employed in bioinformatics data analysis.

Accurate inference of orthologous genes is a pre-requisite for most comparative genomics studies, and is also important for functional annotation of new genomes. Identification of orthologous gene sets typically involves phylogenetic tree analysis, heuristic algorithms based on sequence conservation, synteny analysis, or some combination of these approaches. The most direct tree-based methods typically rely on the comparison of an individual gene tree with a species tree. Once the two trees are accurately constructed, orthologs are straightforwardly identified by the definition of orthology as those homologs that are related by speciation, rather than gene duplication, at their most recent point of origin. Although ideal for the purpose of orthology identification in principle, phylogenetic trees are computationally expensive to construct for large numbers of genes and genomes, and they often contain errors, especially at large evolutionary distances. Moreover, in many organisms, in particular prokaryotes and viruses, evolution does not appear to have followed a simple ‘tree-like’ mode, which makes conventional tree reconciliation inapplicable. Other, heuristic methods identify probable orthologs as the closest homologous pairs or groups of genes in a set of organisms. These approaches are faster and easier to automate than tree-based methods, with efficient implementations provided by graph-theoretical algorithms enabling comparisons of thousands of genomes. Comparisons of these two approaches show that, despite conceptual differences, they produce similar sets of orthologs, especially at short evolutionary distances. Synteny also can aid in identification of orthologs. Often, tree-based, sequence similarity- and synteny-based approaches can be combined into flexible hybrid methods.

Finding new uses for existing drugs, or drug repositioning, has been used as a strategy for decades to get drugs to more patients. As the ability to measure molecules in high-throughput ways has improved over the past decade, it is logical that such data might be useful for enabling drug repositioning through computational methods. Many computational predictions for new indications have been borne out in cellular model systems, though extensive animal model and clinical trial-based validation are still pending. In this review, we show that computational methods for drug repositioning can be classified in two axes: drug based, where discovery initiates from the chemical perspective, or disease based, where discovery initiates from the clinical perspective of disease or its pathology. Newer algorithms for computational drug repositioning will likely span these two axes, will take advantage of newer types of molecular measurements, and will certainly play a role in reducing the global burden of disease.

A recent study examined the stability of rankings from random forests using two variable importance measures (mean decrease accuracy (MDA) and mean decrease Gini (MDG)) and concluded that rankings based on the MDG were more robust than MDA. However, studies examining data-specific characteristics on ranking stability have been few. Rankings based on the MDG measure showed sensitivity to within-predictor correlation and differences in category frequencies, even when the number of categories was held constant, and thus may produce spurious results. The MDA measure was robust to these data characteristics. Further, under strong within-predictor correlation, MDG rankings were less stable than those using MDA.

Proposed molecular classifiers may be overfit to idiosyncrasies of noisy genomic and proteomic data. Cross-validation methods are often used to obtain estimates of classification accuracy, but both simulations and case studies suggest that, when inappropriate methods are used, bias may ensue. Bias can be bypassed and generalizability can be tested by external (independent) validation. We evaluated 35 studies that have reported on external validation of a molecular classifier. We extracted information on study design and methodological features, and compared the performance of molecular classifiers in internal cross-validation versus external validation for 28 studies where both had been performed. We demonstrate that the majority of studies pursued cross-validation practices that are likely to overestimate classifier performance. Most studies were markedly underpowered to detect a 20% decrease in sensitivity or specificity between internal cross-validation and external validation [median power was 36% (IQR, 21–61%) and 29% (IQR, 15–65%), respectively]. The median reported classification performance for sensitivity and specificity was 94% and 98%, respectively, in cross-validation and 88% and 81% for independent validation. The relative diagnostic odds ratio was 3.26 (95% CI 2.04–5.21) for cross-validation versus independent validation. Finally, we reviewed all studies (n = 758) which cited those in our study sample, and identified only one instance of additional subsequent independent validation of these classifiers. In conclusion, these results document that many cross-validation practices employed in the literature are potentially biased and genuine progress in this field will require adoption of routine external validation of molecular classifiers, preferably in much larger studies than in current practice.

Developments in whole genome biotechnology have stimulated statistical focus on prediction methods. We review here methodology for classifying patients into survival risk groups and for using cross-validation to evaluate such classifications. Measures of discrimination for survival risk models include separation of survival curves, time-dependent ROC curves and Harrell’s concordance index. For high-dimensional data applications, however, computing these measures as re-substitution statistics on the same data used for model development results in highly biased estimates. Most developments in methodology for survival risk modeling with high-dimensional data have utilized separate test data sets for model evaluation. Cross-validation has sometimes been used for optimization of tuning parameters. In many applications, however, the data available are too limited for effective division into training and test sets and consequently authors have often either reported re-substitution statistics or analyzed their data using binary classification methods in order to utilize familiar cross-validation. In this article we have tried to indicate how to utilize cross-validation for the evaluation of survival risk models; specifically how to compute cross-validated estimates of survival distributions for predicted risk groups and how to compute cross-validated time-dependent ROC curves. We have also discussed evaluation of the statistical significance of a survival risk model and evaluation of whether high-dimensional genomic data adds predictive accuracy to a model based on standard covariates alone.

Funding bodies are increasingly recognizing the need to provide graduates and researchers with access to short intensive courses in a variety of disciplines, in order both to improve the general skills base and to provide solid foundations on which researchers may build their careers. In response to the development of ‘high-throughput biology’, the need for training in the field of bioinformatics, in particular, is seeing a resurgence: it has been defined as a key priority by many Institutions and research programmes and is now an important component of many grant proposals. Nevertheless, when it comes to planning and preparing to meet such training needs, tension arises between the reward structures that predominate in the scientific community which compel individuals to publish or perish, and the time that must be devoted to the design, delivery and maintenance of high-quality training materials. Conversely, there is much relevant teaching material and training expertise available worldwide that, were it properly organized, could be exploited by anyone who needs to provide training or needs to set up a new course. To do this, however, the materials would have to be centralized in a database and clearly tagged in relation to target audiences, learning objectives, etc. Ideally, they would also be peer reviewed, and easily and efficiently accessible for downloading. Here, we present the Bioinformatics Training Network (BTN), a new enterprise that has been initiated to address these needs and review it, respectively, to similar initiatives and collections.

Advanced research requires intensive interaction among a multitude of actors, often possessing different expertise and usually working at a distance from each other. The field of collaborative research aims to establish suitable models and technologies to properly support these interactions. In this article, we first present the reasons for an interest of Bioinformatics in this context by also suggesting some research domains that could benefit from collaborative research. We then review the principles and some of the most relevant applications of social networking, with a special attention to networks supporting scientific collaboration, by also highlighting some critical issues, such as identification of users and standardization of formats. We then introduce some systems for collaborative document creation, including wiki systems and tools for ontology development, and review some of the most interesting biological wikis. We also review the principles of Collaborative Development Environments for software and show some examples in Bioinformatics. Finally, we present the principles and some examples of Learning Management Systems. In conclusion, we try to devise some of the goals to be achieved in the short term for the exploitation of these technologies.

Biomedical research relies increasingly on large collections of data sets and knowledge whose generation, representation and analysis often require large collaborative and interdisciplinary efforts. This dimension of ‘big data’ research calls for the development of computational tools to manage such a vast amount of data, as well as tools that can improve communication and access to information from collaborating researchers and from the wider community. Whenever research projects have a defined temporal scope, an additional issue of data management arises, namely how the knowledge generated within the project can be made available beyond its boundaries and life-time. DC-THERA is a European ‘Network of Excellence’ (NoE) that spawned a very large collaborative and interdisciplinary research community, focusing on the development of novel immunotherapies derived from fundamental research in dendritic cell immunobiology. In this article we introduce the DC-THERA Directory, which is an information system designed to support knowledge management for this research community and beyond. We present how the use of metadata and Semantic Web technologies can effectively help to organize the knowledge generated by modern collaborative research, how these technologies can enable effective data management solutions during and beyond the project lifecycle, and how resources such as the DC-THERA Directory fit into the larger context of e-science.

The goal of the Gene Ontology (GO) project is to provide a uniform way to describe the functions of gene products from organisms across all kingdoms of life and thereby enable analysis of genomic data. Protein annotations are either based on experiments or predicted from protein sequences. Since most sequences have not been experimentally characterized, most available annotations need to be based on predictions. To make as accurate inferences as possible, the GO Consortium's Reference Genome Project is using an explicit evolutionary framework to infer annotations of proteins from a broad set of genomes from experimental annotations in a semi-automated manner. Most components in the pipeline, such as selection of sequences, building multiple sequence alignments and phylogenetic trees, retrieving experimental annotations and depositing inferred annotations, are fully automated. However, the most crucial step in our pipeline relies on software-assisted curation by an expert biologist. This curation tool, Phylogenetic Annotation and INference Tool (PAINT) helps curators to infer annotations among members of a protein family. PAINT allows curators to make precise assertions as to when functions were gained and lost during evolution and record the evidence (e.g. experimentally supported GO annotations and phylogenetic information including orthology) for those assertions. In this article, we describe how we use PAINT to infer protein function in a phylogenetic context with emphasis on its strengths, limitations and guidelines. We also discuss specific examples showing how PAINT annotations compare with those generated by other highly used homology-based methods.

In recent years, biological web resources such as databases and tools have become more complex because of the enormous amounts of data generated in the field of life sciences. Traditional methods of distributing tutorials include publishing textbooks and posting web documents, but these static contents cannot adequately describe recent dynamic web services. Due to improvements in computer technology, it is now possible to create dynamic content such as video with minimal effort and low cost on most modern computers. The ease of creating and distributing video tutorials instead of static content improves accessibility for researchers, annotators and curators. This article focuses on online video repositories for educational and tutorial videos provided by resource developers and users. It also describes a project in Japan named TogoTV (http://togotv.dbcls.jp/en/) and discusses the production and distribution of high-quality tutorial videos, which would be useful to viewer, with examples. This article intends to stimulate and encourage researchers who develop and use databases and tools to distribute how-to videos as a tool to enhance product usability.

During a meeting of the SYSGENET working group ‘Bioinformatics’, currently available software tools and databases for systems genetics in mice were reviewed and the needs for future developments discussed. The group evaluated interoperability and performed initial feasibility studies. To aid future compatibility of software and exchange of already developed software modules, a strong recommendation was made by the group to integrate HAPPY and R/qtl analysis toolboxes, GeneNetwork and XGAP database platforms, and TIQS and xQTL processing platforms. R should be used as the principal computer language for QTL data analysis in all platforms and a ‘cloud’ should be used for software dissemination to the community. Furthermore, the working group recommended that all data models and software source code should be made visible in public repositories to allow a coordinated effort on the use of common data structures and file formats.

Phylogenomic databases provide orthology predictions for species with fully sequenced genomes. Although the goal seems well-defined, the content of these databases differs greatly. Seven ortholog databases (Ensembl Compara, eggNOG, HOGENOM, InParanoid, OMA, OrthoDB, Panther) were compared on the basis of reference trees. For three well-conserved protein families, we observed a generally high specificity of orthology assignments for these databases. We show that differences in the completeness of predicted gene relationships and in the phylogenetic information are, for the great majority, not due to the methods used, but to differences in the underlying database concepts. According to our metrics, none of the databases provides a fully correct and comprehensive protein classification. Our results provide a framework for meaningful and systematic comparisons of phylogenomic databases. In the future, a sustainable set of ‘Gold standard’ phylogenetic trees could provide a robust method for phylogenomic databases to assess their current quality status, measure changes following new database releases and diagnose improvements subsequent to an upgrade of the analysis procedure.

Ortholog identification is used in gene functional annotation, species phylogeny estimation, phylogenetic profile construction and many other analyses. Bioinformatics methods for ortholog identification are commonly based on pairwise protein sequence comparisons between whole genomes. Phylogenetic methods of ortholog identification have also been developed; these methods can be applied to protein data sets sharing a common domain architecture or which share a single functional domain but differ outside this region of homology. While promiscuous domains represent a challenge to all orthology prediction methods, overall structural similarity is highly correlated with proximity in a phylogenetic tree, conferring a degree of robustness to phylogenetic methods. In this article, we review the issues involved in orthology prediction when data sets include sequences with structurally heterogeneous domain architectures, with particular attention to automated methods designed for high-throughput application, and present a case study to illustrate the challenges in this area.

Recent development of deep sequencing technologies has facilitated de novo genome sequencing projects, now conducted even by individual laboratories. However, this will yield more and more genome sequences that are not well assembled, and will hinder thorough annotation when no closely related reference genome is available. One of the challenging issues is the identification of protein-coding sequences split into multiple unassembled genomic segments, which can confound orthology assignment and various laboratory experiments requiring the identification of individual genes. In this study, using the genome of a cartilaginous fish, Callorhinchus milii, as test case, we performed gene prediction using a model specifically trained for this genome. We implemented an algorithm, designated ESPRIT, to identify possible linkages between multiple protein-coding portions derived from a single genomic locus split into multiple unassembled genomic segments. We developed a validation framework based on an artificially fragmented human genome, improvements between early and recent mouse genome assemblies, comparison with experimentally validated sequences from GenBank, and phylogenetic analyses. Our strategy provided insights into practical solutions for efficient annotation of only partially sequenced (low-coverage) genomes. To our knowledge, our study is the first formulation of a method to link unassembled genomic segments based on proteomes of relatively distantly related species as references.

Orthology is a powerful refinement of homology that allows us to describe more precisely the evolution of genomes and understand the function of the genes they contain. However, because orthology is not concerned with genomic position, it is limited in its ability to describe genes that are likely to have equivalent roles in different genomes. Because of this limitation, the concept of ‘positional orthology’ has emerged, which describes the relation between orthologous genes that retain their ancestral genomic positions. In this review, we formally define this concept, for which we introduce the shorter term ‘toporthology’, with respect to the evolutionary events experienced by a gene’s ancestors. Through a discussion of recent studies on the role of genomic context in gene evolution, we show that the distinction between orthology and toporthology is biologically significant. We then review a number of orthology prediction methods that take genomic context into account and thus that may be used to infer the important relation of toporthology.

Despite the common assumption that orthologs usually share the same function, there have been various reports of divergence between orthologs, even among species as close as mammals. The comparison of mouse and human is of special interest, because mouse is often used as a model organism to understand human biology. We review the literature on evidence for divergence between human and mouse orthologous genes, and discuss it in the context of biomedical research.