Search

We are amidst an ongoing flood of sequence data arising from the application of high-throughput technologies, and a concomitant fundamental revision in our understanding of how genomes evolve individually and within the biosphere. Workflows for phylogenomic inference must accommodate data that are not only much larger than before, but often more error prone and perhaps misassembled, or not assembled in the first place. Moreover, genomes of microbes, viruses and plasmids evolve not only by tree-like descent with modification but also by incorporating stretches of exogenous DNA. Thus, next-generation phylogenomics must address computational scalability while rethinking the nature of orthogroups, the alignment of multiple sequences and the inference and comparison of trees. New phylogenomic workflows have begun to take shape based on so-called alignment-free (AF) approaches. Here, we review the conceptual foundations of AF phylogenetics for the hierarchical (vertical) and reticulate (lateral) components of genome evolution, focusing on methods based on k-mers. We reflect on what seems to be successful, and on where further development is needed.

16 páginas, 1 tabla, 3 figuras., Mesenteric infection by the parasitic blood fluke Schistosoma bovis is a common veterinary problem in Africa and the Middle East and occasionally in the Mediterranean Region. The species also has the ability to form interspecific hybrids with the human parasite S. haematobium with natural hybridisation observed in West Africa, presenting possible zoonotic transmission. Additionally, this exchange of alleles between species may dramatically influence disease dynamics and parasite evolution. We have generated a 374 Mb assembly of the S. bovis genome using Illumina and PacBio-based technologies. Despite infecting different hosts and organs, the genome sequences of S. bovis and S. haematobium appeared strikingly similar with 97% sequence identity. The two species share 98% of protein-coding genes, with an average sequence identity of 97.3% at the amino acid level. Genome comparison identified large continuous parts of the genome (up to several 100 kb) showing almost 100% sequence identity between S. bovis and S. haematobium. It is unlikely that this is a result of genome conservation and provides further evidence of natural interspecific hybridization between S. bovis and S. haematobium. Our results suggest that foreign DNA obtained by interspecific hybridization was maintained in the population through multiple meiosis cycles and that hybrids were sexually reproductive, producing viable offspring. The S. bovis genome assembly forms a highly valuable resource for studying schistosome evolution and exploring genetic regions that are associated with species-specific phenotypic traits., This work was supported by grants from the QIMR Berghofer Medical Research Institute (Chenhall Estate) and the Australian Infectious Diseases Research Centre. NDY was supported by a National Health and Medical Research Council (NHMRC) CDF1, PKK by an NHMRC ECRF and DPM by an NHMRC SPRF (APP1102926). The work has been supported by NHMRC project grant APP1102322.

Genome-wide regulatory networks enable cells to function, develop, and survive. Perturbation of these networks can lead to appearance of a disease phenotype. Inspired by Conrad Waddington's epigenetic landscape of cell development, we use a Hopfield network formalism to construct an attractor landscape model of disease progression based on protein- or gene-correlation networks of Parkinson's disease, glioma, and colorectal cancer. Attractors in this landscape correspond to normal and disease states of the cell. We introduce approaches to estimate the size and robustness of these attractors, and take a network-based approach to study their biological features such as the key genes and their functions associated with the attractors. Our results show that the attractor of cancer cells is wider than the attractor of normal cells, suggesting a heterogeneous nature of cancer. Perturbation analysis shows that robustness depends on characteristics of the input data (number of samples per time-point, and the fraction which converge to an attractor). We identify unique gene interactions at each stage, which reflect the temporal rewiring of the gene regulatory network (GRN) with disease progression. Our model of the attractor landscape, constructed from large-scale gene expression profiles of individual patients, captures snapshots of disease progression and identifies gene interactions specific to different stages, opening the way for development of stage-specific therapeutic strategies.

Accurate subtyping or classification of breast cancer is important for ensuring proper treatment of patients and also for understanding the molecular mechanisms driving this disease. While there have been several gene signatures proposed in the literature to classify breast tumours, these signatures show very low overlaps, different classification performance, and not much relevance to the underlying biology of these tumours. Here we evaluate DNA-damage response (DDR) and cell cycle pathways, which are critical pathways implicated in a considerable proportion of breast tumours, for their usefulness and ability in breast tumour subtyping. We think that subtyping breast tumours based on these two pathways could lead to vital insights into molecular mechanisms driving these tumours. Here, we performed a systematic evaluation of DDR and cell-cycle pathways for subtyping of breast tumours into the five known intrinsic subtypes. Homologous Recombination (HR) pathway showed the best performance in subtyping breast tumours, indicating that HR genes are strongly involved in all breast tumours. Comparisons of pathway based signatures and two standard gene signatures supported the use of known pathways for breast tumour subtyping. Further, the evaluation of these standard gene signatures showed that breast tumour subtyping, prognosis and survival estimation are all closely related. Finally, we constructed an all-inclusive super-signature by combining (union of) all genes and performing a stringent feature selection, and found it to be reasonably accurate and robust in classification as well as prognostic value. Adopting DDR and cell cycle pathways for breast tumour subtyping achieved robust and accurate breast tumour subtyping, and constructing a super-signature which contains feature selected mix of genes from these molecular pathways as well as clinical aspects is valuable in clinical practice.

From 1971 to 1985, Carl Woese and colleagues generated oligonucleotide catalogs of 16S/18S rRNAs from more than 400 organisms. Using these incomplete and imperfect data, Carl and his colleagues developed unprecedented insights into the structure, function, and evolution of the large RNA components of the translational apparatus. They recognized a third domain of life, revealed the phylogenetic backbone of bacteria (and its limitations), delineated taxa, and explored the tempo and mode of microbial evolution. For these discoveries to have stood the test of time, oligonucleotide catalogs must carry significant phylogenetic signal; they thus bear re-examination in view of the current interest in alignment-free phylogenetics based on k-mers. Here we consider the aims, successes, and limitations of this early phase of molecular phylogenetics. We computationally generate oligonucleotide sets (e-catalogs) from 16S/18S rRNA sequences, calculate pairwise distances between them based on D 2 statistics, compute distance trees, and compare their performance against alignment-based and k-mer trees. Although the catalogs themselves were superseded by full-length sequences, this stage in the development of computational molecular biology remains instructive for us today.

Large quantities of information describing the mechanisms of biological pathways continue to be collected in publicly available databases. At the same time, experiments have increased in scale, and biologists increasingly use pathways defined in online databases to interpret the results of experiments and generate hypotheses. Emerging computational techniques that exploit the rich biological information captured in reaction systems require formal standardized descriptions of pathways to extract these reaction networks and avoid the alternative: time-consuming and largely manual literature-based network reconstruction. Here, we systematically evaluate the effects of commonly used knowledge representations on the seemingly simple task of extracting a reaction network describing signal transduction from a pathway database. We show that this process is in fact surprisingly difficult, and the pathway representations adopted by various knowledge bases have dramatic consequences for reaction network extraction, connectivity, capture of pathway crosstalk and in the modelling of cell-cell interactions. Researchers constructing computational models built from automatically extracted reaction networks must therefore consider the issues we outline in this review to maximize the value of existing pathway knowledge.

Recent advances in molecular technologies have opened up unprecedented opportunities for molecular ecologists to better understand the molecular basis of traits of ecological and evolutionary importance in almost any organism. Nevertheless, reliable and systematic inference of functionally relevant information from these masses of data remains challenging. The aim of this review is to highlight how the Gene Ontology (GO) database can be of use in resolving this challenge. The GO provides a largely species-neutral source of information on the molecular function, biological role and cellular location of tens of thousands of gene products. As it is designed to be species-neutral, the GO is well suited for cross-species use, meaning that, functional annotation derived from model organisms can be transferred to inferred orthologues in newly sequenced species. In other words, the GO can provide gene annotation information for species with nonannotated genomes. In this review, we describe the GO database, how functional information is linked with genes/gene products in model organisms, and how molecular ecologists can utilize this information to annotate their own data. Then, we outline various applications of GO for enhancing the understanding of molecular basis of traits in ecologically relevant species. We also highlight potential pitfalls, provide step-by-step recommendations for conducting a sound study in nonmodel organisms, suggest avenues for future research and outline a strategy for maximizing the benefits of a more ecological and evolutionary genomics-oriented ontology by ensuring its compatibility with the GO.

Bacteria and archaea can exchange genetic material across lineages through processes of lateral genetic transfer (LGT). Collectively, these exchange relationships can be modeled as a network and analyzed using concepts from graph theory. In particular, densely connected regions within an LGT network have been defined as genetic exchange communities (GECs). However, it has been problematic to construct networks in which edges solely represent LGT. Here we apply term frequency-inverse document frequency (TF-IDF), an alignment-free method originating from document analysis, to infer regions of lateral origin in bacterial genomes. We examine four empirical datasets of different size (number of genomes) and phyletic breadth, varying a key parameter (word length k) within bounds established in previous work. We map the inferred lateral regions to genes in recipient genomes, and construct networks in which the nodes are groups of genomes, and the edges natively represent LGT. We then extract maximum and maximal cliques (i.e., GECs) from these graphs, and identify nodes that belong to GECs across a wide range of k. Most surviving lateral transfer has happened within these GECs. Using Gene Ontology enrichment tests we demonstrate that biological processes associated with metabolism, regulation and transport are often over-represented among the genes affected by LGT within these communities. These enrichments are largely robust to change of k.

Lateral genetic transfer (LGT) is the process by which genetic material moves between organisms (and viruses) in the biosphere. Among the many approaches developed for the inference of LGT events from DNA sequence data, methods based on the comparison of phylogenetic trees remain the gold standard for many types of problem. Identifying LGT events from sequenced genomes typically involves a series of steps in which homologous sequences are identified and aligned, phylogenetic trees are inferred, and their topologies are compared to identify unexpected or conflicting relationships. These types of approach have been used to elucidate the nature and extent of LGT and its physiological and ecological consequences throughout the Tree of Life. Advances in DNA sequencing technology have led to enormous increases in the number of sequenced genomes, including ultra-deep sampling of specific taxonomic groups and single cell-based sequencing of unculturable “microbial dark matter.” Environmental shotgun sequencing enables the study of LGT among organisms that share the same habitat. This abundance of genomic data offers new opportunities for scientific discovery, but poses two key problems. As ever more genomes are generated, the assembly and annotation of each individual genome receives less scrutiny; and with so many genomes available it is tempting to include them all in a single analysis, but thousands of genomes and millions of genes can overwhelm key algorithms in the analysis pipeline. Identifying LGT events of interest therefore depends on choosing the right dataset, and on algorithms that appropriately balance speed and accuracy given the size and composition of the chosen set of genomes.

Lateral genetic transfer (LGT) plays an important role in the evolution of microbes. Existing computational methods for detecting genomic regions of putative lateral origin scale poorly to large data. Here, we propose a novel method based on TF-IDF (Term Frequency-Inverse Document Frequency) statistics to detect not only regions of lateral origin, but also their origin and direction of transfer, in sets of hierarchically structured nucleotide or protein sequences. This approach is based on the frequency distributions of k-mers in the sequences. If a set of contiguous k-mers appears sufficiently more frequently in another phyletic group than in its own, we infer that they have been transferred from the first group to the second. We performed rigorous tests of TF-IDF using simulated and empirical datasets. With the simulated data, we tested our method under different parameter settings for sequence length, substitution rate between and within groups and post-LGT, deletion rate, length of transferred region and k size and found that we can detect LGT events with high precision and recall. Our method performs better than an established method, ALFY, which has high recall but low precision. Our method is efficient, with runtime increasing approximately linearly with sequence length.

Many microbes can acquire genetic material from their environment and incorporate it into their genome, a process known as lateral genetic transfer (LGT). Computational approaches have been developed to detect genomic regions of lateral origin, but typically lack sensitivity, ability to distinguish donor from recipient and scalability to very large datasets. To address these issues we have introduced an alignment-free method based on ideas from document analysis, term frequency-inverse document frequency (TF-IDF). Here we examine the performance of TF-IDF on three empirical datasets: 27 genomes of Escherichia coli and Shigella, 110 genomes of enteric bacteria and 143 genomes across 12 bacterial and three archaeal phyla. We investigate the effect of k-mer size, gap size and delineation of groups on the inference of genomic regions of lateral origin, finding an interplay among these parameters and sequence divergence. Because TF-IDF identifies donor groups and delineates regions of lateral origin within recipient genomes, aggregating these regions by gene enables us to explore, for the first time, the mosaic nature of lateral genes including the multiplicity of biological sources, ancestry of transfer and over-writing by subsequent transfers. We carry out Gene Ontology enrichment tests to investigate which biological processes are potentially affected by LGT.

Alignment-free (AF) approaches have recently been highlighted as alternatives to methods based on multiple sequence alignment in phylogenetic inference. However, the sensitivity of AF methods to genome-scale evolutionary scenarios is little known. Here, using simulated microbial genome data we systematically assess the sensitivity of nine AF methods to three important evolutionary scenarios: sequence divergence, lateral genetic transfer (LGT) and genome rearrangement. Among these, AF methods are most sensitive to the extent of sequence divergence, less sensitive to low and moderate frequencies of LGT and most robust against genome rearrangement. We describe the application of AF methods to three well-studied empirical genome datasets and introduce a new application of the jackknife to assess node support. Our results demonstrate that AF phylogenomics is computationally scalable to multi-genome data and can generate biologically meaningful phylogenies and insights into microbial evolution.

The widespread presence of antibiotic resistance and virulence among Staphylococcus isolates has been attributed in part to lateral genetic transfer (LGT), but little is known about the broader extent of LGT within this genus. Here we report the first systematic study of the modularity of genetic transfer among 13 Staphylococcus genomes covering four distinct named species. Using a topology-based phylogenetic approach, we found, among 1,354 sets of homologous genes examined, strong evidence of LGT in 368 (27.1%) gene sets, and weaker evidence in another 259 (19.1%). Within-gene and whole-gene transfer contribute almost equally to the topological discordance of these gene sets against a reference phylogeny. Comparing genetic transfer in single-copy and in multicopy gene sets, we observed a higher frequency of LGT in the latter, and a substantial functional bias in cases of whole-gene transfer (little such bias was observed in cases of fragmentary genetic transfer). We found evidence that lateral transfer, particularly of entire genes, impacts not only functions related to antibiotic, drug, and heavy-metal resistance, as well as membrane transport, but also core informational and metabolic functions not associated with mobile elements. Although patterns of sequence similarity support the cohesion of recognized species, LGT within S. aureus appears frequently to disrupt clonal complexes. Our results demonstrate that LGT and gene duplication play important parts in functional innovation in staphylococcal genomes.

A wide range of data on sequences, structures, pathways, and networks of genes and gene products is available for hypothesis testing and discovery in biological and biomedical research. However, data describing the physical, chemical, and biological properties of small molecules have not been well-integrated with these resources. Semantically rich representations of chemical data, combined with Semantic Web technologies, have the potential to enable the integration of small molecule and biomolecular data resources, expanding the scope and power of biomedical and pharmacological research. We employed the Semantic Web technologies Resource Description Framework (RDF) and Web Ontology Language (OWL) to generate a Small Molecule Ontology (SMO) that represents concepts and provides unique identifiers for biologically relevant properties of small molecules and their interactions with biomolecules, such as proteins. We instanced SMO using data from three public data sources, i.e., DrugBank, PubChem and UniProt, and converted to RDF triples. Evaluation of SMO by use of predetermined competency questions implemented as SPARQL queries demonstrated that data from chemical and biomolecular data sources were effectively represented and that useful knowledge can be extracted. These results illustrate the potential of Semantic Web technologies in chemical, biological, and pharmacological research and in drug discovery.

Protein-protein interactions (PPIs) underpin key biological processes in cells. Describing and interpreting this network of interactions is a key focus of computational systems biology. While mouse is commonly used as a model system for mammalian biology, the description of mouse PPIs available in public interaction databases is remarkably poor. Collectively, public resources such as BIND, IntACT and MINT contain only a few thousand mouse PPIs, far behind the many tens or hundreds of thousands likely to exist. To supplement this lack and to take advantage of other high-throughput omic data sets in mouse, here we identify that portion of the human interactome with orthologs in mouse, and from that infer a mouse interolog network. By inferring interactions in mouse based on only the most closely related species with abundant PPI data (human), we create a view of mouse interactions enriched for shared mammalian biological processes. We also demonstrate that available methods for determining orthologs between even closely related species produce distinctly different results, and we propose an integrated view of mouse-human orthology from which to infer a broader interolog network.

The significance of whole-genome duplications (WGD) for vertebrate evolution remains controversial, in part because the mechanisms by which WGD contributed to functional evolution or speciation are still incompletely characterized. Fish genomes provide an ideal context in which to examine the consequences of WGD, because the teleost lineage experienced an additional WGD soon after divergence from tetrapods and because five teleost genomes are available for comparative analysis. Here we present an integrated approach to characterize these post-duplication genomes based on genome-scale synteny, phylogenetic, temporal, and spatial gene expression and on protein sequence data. A minimum of 3%–4% of protein-coding loci have been retained in two copies in each of the five fish genomes, and many of these duplicates are key developmental genes that function as transcription factors or signaling molecules. Almost all duplicate gene pairs we examined have diverged in spatial and/or temporal expression during embryogenesis. A quarter of duplicate pairs have diverged in function via the acquisition of novel protein domains or via changes in the subcellular localization of their encoded proteins. We compared the spatial expression and protein domain architecture of zebrafish WGD-duplicates to those of their single mouse ortholog and found many examples supporting a model of neofunctionalization. WGD-duplicates have acquired novel protein domains more often than have single-copy genes. Post-WGD changes at the gene regulatory level were more common than changes at the protein level. We conclude that the most significant consequence of WGD for vertebrate evolution has been to enable more-specialized regulatory control of development via the acquisition of novel spatiotemporal expression domains. We find limited evidence that reciprocal gene loss led to reproductive isolation and speciation in this lineage.

Pairwise local sequence alignment methods have been the prevailing technique to identify homologous nucleotides between related species. However, existing methods that identify and align all homologous nucleotides in one or more genomes have suffered from poor scalability and limited accuracy. We propose a novel method that couples a gapped extension heuristic with an efficient filtration method for identifying interspersed repeats in genome sequences. During gapped extension, we use the MUSCLE implementation of progressive global multiple alignment with iterative refinement. The resulting gapped extensions potentially contain alignments of unrelated sequence. We detect and remove such undesirable alignments using a hidden Markov model (HMM) to predict the posterior probability of homology. The HMM emission frequencies for nucleotide substitutions can be derived from any time-reversible nucleotide substitution matrix. We evaluate the performance of our method and previous approaches on a hybrid data set of real genomic DNA with simulated interspersed repeats. Our method outperforms a related method in terms of sensitivity, positive predictive value, and localizing boundaries of homology. The described methods have been implemented in freely available software, Repeatoire, available from: http://wwwabi.snv.jussieu.fr/public/Repeatoire. © 2006 IEEE.

The epigenetic landscape was introduced by Conrad Waddington as a metaphor of cellular development. Like a ball rolling down a hillside is channelled through a succession of valleys until it reaches the bottom, cells follow specific trajectories from a pluripotent state to a committed state. Transcription factors (TFs) interacting as a network (the gene regulatory network (GRN)) orchestrate this developmental process within each cell. Here, we quantitatively model the epigenetic landscape using a kind of artificial neural network called the Hopfield network (HN). An HN is composed of nodes (genes/TFs) and weighted undirected edges, resulting in a weight matrix (W) that stores interactions among the nodes over the entire network. We used gene co-expression to compute the edge weights. Through W, we then associate an energy score (E) to each input pattern (pattern of co-expression for a specific developmental stage) such that each pattern has a specific E. We propose that, based on the co-expression values stored in W, HN associates lower E values to stable phenotypic states and higher E to transient states. We validate our model using time course gene-expression data sets representing stages of development across 12 biological processes including differentiation of human embryonic stem cells into specialized cells, differentiation of THP1 monocytes to macrophages during immune response and trans-differentiation of epithelial to mesenchymal cells in cancer. We observe that transient states have higher energy than the stable phenotypic states, yielding an arc-shaped trajectory. This relationship was confirmed by perturbation analysis. HNs offer an attractive framework for quantitative modelling of cell differentiation (as a landscape) from empirical data. Using HNs, we identify genes and TFs that drive cell-fate transitions, and gain insight into the global dynamics of GRNs.

The process of inferring phylogenetic trees from molecular sequences almost always starts with a multiple alignment of these sequences but can also be based on methods that do not involve multiple sequence alignment. Very little is known about the accuracy with which such alignment-free methods recover the correct phylogeny or about the potential for increasing their accuracy. We conducted a large-scale comparison of ten alignment-free methods, among them one new approach that does not calculate distances and a faster variant of our pattern-based approach; all distance-based alignment-free methods are freely available from http://www.bioinformatics.org.au (as Python package decaf+py). We show that most methods exhibit a higher overall reconstruction accuracy in the presence of high among-site rate variation. Under all conditions that we considered, variants of the pattern-based approach were significantly better than the other alignment-free methods. The new pattern-based variant achieved a speed-up of an order of magnitude in the distance calculation step, accompanied by a small loss of tree reconstruction accuracy. A method of Bayesian inference from k-mers did not improve on classical alignment-free (and distance-based) methods but may still offer other advantages due to its Bayesian nature. We found the optimal word length k of word-based methods to be stable across various data sets, and we provide parameter ranges for two different alphabets. The influence of these alphabets was analyzed to reveal a trade-off in reconstruction accuracy between long and short branches. We have mapped the phylogenetic accuracy for many alignment-free methods, among them several recently introduced ones, and increased our understanding of their behavior in response to biologically important parameters. In all experiments, the pattern-based approach emerged as superior, at the expense of higher resource consumption. Nonetheless, no alignment-free method that we examined recovers the correct phylogeny as accurately as does an approach based on maximum-likelihood distance estimates of multiply aligned sequences.

A frequent bottleneck in interpreting phylogenomic output is the need to screen often thousands of trees for features of interest, particularly robust clades of specific taxa, as evidence of monophyletic relationship and/or reticulated evolution. Here we present PhySortR, a fast, flexible R package for classifying phylogenetic trees. Unlike existing utilities, PhySortR allows for identification of both exclusive and non-exclusive clades uniting the target taxa based on tip labels (i.e., leaves) on a tree, with customisable options to assess clades within the context of the whole tree. Using simulated and empirical datasets, we demonstrate the potential and scalability of PhySortR in analysis of thousands of phylogenetic trees without a priori assumption of tree-rooting, and in yielding readily interpretable trees that unambiguously satisfy the query. PhySortR is a command-line tool that is freely available and easily automatable.

Background: Synthetic lethality (SL) refers to the genetic interaction between two or more genes where only their co-alteration (e.g. by mutations, amplifications or deletions) results in cell death. In recent years, SL has emerged as an attractive therapeutic strategy against cancer: by targeting the SL partners of altered genes in cancer cells, these cells can be selectively killed while sparing the normal cells. Consequently, a number of studies have attempted prediction of SL interactions in human, a majority by extrapolating SL interactions inferred through large-scale screens in model organisms. However, these predicted SL interactions either do not hold in human cells or do not include genes that are (frequently) altered in human cancers, and are therefore not attractive in the context of cancer therapy. Results: Here, we develop a computational approach to infer SL interactions directly from frequently altered genes in human cancers. It is based on the observation that pairs of genes that are altered in a (significantly) mutually exclusive manner in cancers are likely to constitute lethal combinations. Using genomic copy-number and gene-expression data from four cancers, breast, prostate, ovarian and uterine (total 3980 samples) from The Cancer Genome Atlas, we identify 718 genes that are frequently amplified or upregulated, and are likely to be synthetic lethal with six key DNA-damage response (DDR) genes in these cancers. By comparing with published data on gene essentiality (~16000 genes) from ten DDR-deficient cancer cell lines, we show that our identified genes are enriched among the top quartile of essential genes in these cell lines, implying that our inferred genes are highly likely to be (synthetic) lethal upon knockdown in these cell lines., 35 pages, 7 figures

The extent to which lateral genetic transfer has shaped microbial genomes has major implications for the emergence of community structures. We have performed a rigorous phylogenetic analysis of >220,000 proteins from genomes of 144 prokaryotes to determine the contribution of gene sharing to current prokaryotic diversity, and to identify “highways” of sharing between lineages. The inferred relationships suggest a pattern of inheritance that is largely vertical, but with notable exceptions among closely related taxa, and among distantly related organisms that live in similar environments.

Integrating information in the molecular biosciences involves more than the cross-referencing of sequences or structures. Experimental protocols, results of computational analyses, annotations and links to relevant literature form integral parts of this information, and impart meaning to sequence or structure. In this review, we examine some existing approaches to integrating information in the molecular biosciences. We consider not only technical issues concerning the integration of heterogeneous data sources and the corresponding semantic implications, but also the integration of analytical results. Within the broad range of strategies for integration of data and information, we distinguish between platforms and developments. We discuss two current platforms and six current developments, and identify what we believe to be their strengths and limitations. We identify key unsolved problems in integrating information in the molecular biosciences, and discuss possible strategies for addressing them including semantic integration using ontologies, XML as a data model, and graphical user interfaces as integrative environments.

If open reading frames (ORFs) have been transmitted primarily by vertical descent, the distributional profile of orthologues of each ORF should be congruent with the organismal tree or a subtree thereof. Distributional patterns not reconciled parsimoniously with tree-like descent and loss are prima facie evidence of lateral gene transfer. Herein, a rigorous criterion for recognizing ORF distributions is described and implemented; it does not require the inference of phylogenetic trees, nor does it assume any specific tree. Because lineage-specific differences in rates of sequence change can also generate unexpected distributional patterns, rate artefacts were controlled for by requiring pairwise matches between ORFs to exceed a rigorous inclusion threshold, but absence of a match was assessed against a more-permissive exclusion threshold. Applying this dual-threshold criterion to cross-domain and cross-phylum distributional patterns for ORFs in 23 bacterial genomes, a relative abundance of ORFs was observed that find a match in exactly seven other bacterial phyla; 94-99% of these ORFs also find matches among the Archaea and/or Eukarya. In the larger (and some smaller) bacterial genomes, ORFs that find matches in exactly one other bacterial phylum are also relatively abundant, but fewer of these have non-bacterial homologues; most of their matches within the Bacteria are to the Proteobacteria and/or Firmicutes, which cannot be sister lineages to all bacteria. ORFs that are neither distributed universally among the Bacteria, nor necessarily shared with topologically adjacent lineages, are preferentially enriched in large bacterial genomes.

Alignment-free methods, in which shared properties of sub-sequences (e.g. identity or match length) are extracted and used to compute a distance matrix, have recently been explored for phylogenetic inference. However, the scalability and robustness of these methods to key evolutionary processes remain to be investigated. Here, using simulated sequence sets of various sizes in both nucleotides and amino acids, we systematically assess the accuracy of phylogenetic inference using an alignment-free approach, based on D2 statistics, under different evolutionary scenarios. We find that compared to a multiple sequence alignment approach, D2 methods are more robust against among-site rate heterogeneity, compositional biases, genetic rearrangements and insertions/deletions, but are more sensitive to recent sequence divergence and sequence truncation. Across diverse empirical datasets, the alignment-free methods perform well for sequences sharing low divergence, at greater computation speed. Our findings provide strong evidence for the scalability and the potential use of alignment-free methods in large-scale phylogenomics.

Summary: New high-resolution approaches for mapping ultrastructure of cells in 3D are leading to unprecedented quantities of spatial data. Here we present Illoura, a software tool for the integrated management, analysis and visualization of these data within a semantic context, and illustrate its capability by analysis of spatial relationships in mammalian beta cells. Availability: http://www.visiblecell.com/illoura Contact: m.ragan@uq.edu.au Supplementary information: Supplementary data are available at Bioinformatics online.

We examine the translated open reading frames (ORFs) of the yeast Saccharomyces cerevisiae, focusing on those that have FASTA matches in phyletically defined sets of completely sequenced genomes. On this basis, we identify archaeal yeast, bacterial yeast, universal yeast, and yeast ORFs that do not have a match in any of nine prokaryote genomes. Similarly, we examine the yeast mitochondrial genome and the subset of the yeast nuclear ORFs identified as being involved in mitochondrial biogenesis. For the yeast ORFs that match one or more ORFs in these prokaryote genomes, we examine the phyletic and functional distributions of these matches as a function of match strength. These results provide genome level insights into the origin of the eukaryotic cell and the origin of mitochondria. More generally, they exemplify how the growing database of prokaryote genome sequences can help us understand eukaryote genomes.

We have designed and implemented a system to carry out cross-genome comparisons of open reading frames (ORFs) from multiple genomes. This implementation includes a genome profiling system that allows us to explore pairwise comparisons at different levels of match similarity and ask biologically motivated queries involving number and identity of ORFs, their function, functional category, distribution in genomes or in biological domains, and statistics on their matches and match families. This analysis required precise definition of new classification terms and concepts. We define the terms genomic signature, summary signature, biologic domain signature, domain class, match level, match family, and extended match family, then use these terms to define concepts, including genomically universal proteins and proteins characteristics of sets of genomes. We initiate an analysis based on automated FASTA (Pearson, 1996) comparison of 22,419 conceptually translated protein sequences from nine microbial genomes.

Recent work has generated whole-cell and whole-process models capable of predicting phenotype in simple organisms. The approaches used are hindered in higher organisms and more-complex cells by a lack of kinetic parameters for reactions and events, and the difficulty of measuring and estimating these. Here, we outline a large, two-process model capable of predicting the effects of gene expression on a signal transduction network. Our method models signal transduction and the processes involved in gene expression as two separate systems, solved iteratively. We show that this approach is sufficient to capture functionally significant behaviour resulting from common network motifs. We further demonstrate that our method is scalable and efficient to the size of the largest signal transduction databases currently available. This approach enables analysis and prediction in the absence of kinetic data, but is itself held back by the lack of detailed large-scale gene expression models. However, research consortia such as ENCODE and FANTOM are rapidly adding to the knowledge of transcriptional regulation, and we anticipate that incorporating this data into our regulatory model could allow the modelling of complex cellular phenomena such as the structured progression seen in cellular differentiation.

Inference of gene regulatory network from expression data is a challenging task. Many methods have been developed to this purpose but a comprehensive evaluation that covers unsupervised, semi-supervised and supervised methods, and provides guidelines for their practical application, is lacking. We performed an extensive evaluation of inference methods on simulated and experimental expression data. The results reveal low prediction accuracies for unsupervised techniques with the notable exception of the Z-SCORE method on knockout data. In all other cases, the supervised approach achieved the highest accuracies and even in a semi-supervised setting with small numbers of only positive samples, outperformed the unsupervised techniques.

Cancer forms a robust system and progresses as stages over time typically with increasing aggressiveness and worsening prognosis. Characterizing these stages and identifying the genes driving transitions between them is critical to understand cancer progression and to develop effective anti-cancer therapies. Here, we propose a novel model of the 'cancer system' as a Boolean state space in which a Boolean network, built from protein interaction and gene-expression data from different stages of cancer, transits between Boolean satisfiability states by "editing" interactions and "flipping" genes. The application of our model (called BoolSpace) on three case studies - pancreatic and breast tumours in human and post spinal-cord injury in rats - reveals valuable insights into the phenomenon of cancer progression. In particular, we notice that several of the genes flipped are serine/threonine kinases which act as natural cellular switches and that different sets of genes are flipped during the initial and final stages indicating a pattern to tumour progression. We hypothesize that robustness of cancer partly stems from "passing of the baton" between genes at different stages, and therefore an effective therapy should target a "cover set" of these genes. A C/C++ implementation of BoolSpace is freely available at: http://www.bioinformatics.org.au/tools-data, Comment: 13 pages, 4 figures, Genome Informatics (GIW) Singapore, December 2013. To appear in IEEE/ACM Trans. Computational Biology and Bioinformatics

Each step in the temporal process of lateral genetic transfer (LGT) presents barriers and biases that, together with contingent processes in the environment, determine the success of transfer events. Appropriate interaction with existing host systems, together with regulated expression of gene products, are particularly important for the long-term establishment of transferred DNA. Large datasets of biological interactions can be abstracted to construct graphical abstractions of cellular networks. Phyletic distributions and other metadata can then be mapped onto these networks to examine hypotheses of network properties that contribute to successful LGT. Here we examine the organising principles that govern the integration of laterally transferred DNA with existing host cellular networks.

Thanks to advances in next-generation technologies, genome sequences are now being generated at breadth (e.g. across environments) and depth (thousands of closely related strains, individuals or samples) unimaginable only a few years ago. Phylogenomics – the study of evolutionary relationships based on comparative analysis of genome-scale data – has so far been developed as industrial-scale molecular phylogenetics, proceeding in the two classical steps: multiple alignment of homologous sequences, followed by inference of a tree (or multiple trees). However, the algorithms typically employed for these steps scale poorly with number of sequences, such that for an increasing number of problems, high-quality phylogenomic analysis is (or soon will be) computationally infeasible. Moreover, next-generation data are often incomplete and error-prone, and analysis may be further complicated by genome rearrangement, gene fusion and deletion, lateral genetic transfer, and transcript variation. Here we argue that next-generation data require next-generation phylogenomics, including so-called alignment-free approaches. Reviewers Reviewed by Mr Alexander Panchin (nominated by Dr Mikhail Gelfand), Dr Eugene Koonin and Prof Peter Gogarten. For the full reviews, please go to the Reviewers’ comments section.

Motivation: Phylogenetic profiling methods can achieve good accuracy in predicting protein–protein interactions, especially in prokaryotes. Recent studies have shown that the choice of reference taxa (RT) is critical for accurate prediction, but with more than 2500 fully sequenced taxa publicly available, identifying the most-informative RT is becoming increasingly difficult. Previous studies on the selection of RT have provided guidelines for manual taxon selection, and for eliminating closely related taxa. However, no general strategy for automatic selection of RT is currently available.Results: We present three novel methods for automating the selection of RT, using machine learning based on known protein–protein interaction networks. One of these methods in particular, Tree-Based Search, yields greatly improved prediction accuracies. We further show that different methods for constituting phylogenetic profiles often require very different RT sets to support high prediction accuracy. Motivation: Phylogenetic profiling methods can achieve good accuracy in predicting protein-protein interactions, especially in prokaryotes. Recent studies have shown that the choice of reference taxa is critical for accurate prediction, but with more than 2,500 fully sequenced taxa publicly available, identifying the most-informative reference taxa is becoming increasingly difficult. Previous studies on the selection of reference taxa have provided guidelines for manual taxon selection, and for eliminating closely related taxa. However, no general strategy for automatic selection of reference taxa is currently available. Results: We present three novel approaches for automating the selection of reference taxa, using machine learning based on known protein-protein interaction networks. One of these approaches in particular, Tree-Based Search, yields greatly improved prediction accuracy. We further show that different methods for constituting phylogenetic profiles often require very different reference taxon sets to support high prediction accuracy. Availability: All software used in the experiments can be found at www.birc.au.dk/software/PPRTO/

Background Altered networks of gene regulation underlie many complex conditions, including cancer. Inferring gene regulatory networks from high-throughput microarray expression data is a fundamental but challenging task in computational systems biology and its translation to genomic medicine. Although diverse computational and statistical approaches have been brought to bear on the gene regulatory network inference problem, their relative strengths and disadvantages remain poorly understood, largely because comparative analyses usually consider only small subsets of methods, use only synthetic data, and/or fail to adopt a common measure of inference quality. Methods We report a comprehensive comparative evaluation of nine state-of-the art gene regulatory network inference methods encompassing the main algorithmic approaches (mutual information, correlation, partial correlation, random forests, support vector machines) using 38 simulated datasets and empirical serous papillary ovarian adenocarcinoma expression-microarray data. We then apply the best-performing method to infer normal and cancer networks. We assess the druggability of the proteins encoded by our predicted target genes using the CancerResource and PharmGKB webtools and databases. Results We observe large differences in the accuracy with which these methods predict the underlying gene regulatory network depending on features of the data, network size, topology, experiment type, and parameter settings. Applying the best-performing method (the supervised method SIRENE) to the serous papillary ovarian adenocarcinoma dataset, we infer and rank regulatory interactions, some previously reported and others novel. For selected novel interactions we propose testable mechanistic models linking gene regulation to cancer. Using network analysis and visualization, we uncover cross-regulation of angiogenesis-specific genes through three key transcription factors in normal and cancer conditions. Druggabilty analysis of proteins encoded by the 10 highest-confidence target genes, and by 15 genes with differential regulation in normal and cancer conditions, reveals 75% to be potential drug targets. Conclusions Our study represents a concrete application of gene regulatory network inference to ovarian cancer, demonstrating the complete cycle of computational systems biology research, from genome-scale data analysis via network inference, evaluation of methods, to the generation of novel testable hypotheses, their prioritization for experimental validation, and discovery of potential drug targets.

Background The use of ontologies to control vocabulary and structure annotation has added value to genome-scale data, and contributed to the capture and re-use of knowledge across research domains. Gene Ontology (GO) is widely used to capture detailed expert knowledge in genomic-scale datasets and as a consequence has grown to contain many terms, making it unwieldy for many applications. To increase its ease of manipulation and efficiency of use, subsets called GO slims are often created by collapsing terms upward into more general, high-level terms relevant to a particular context. Creation of a GO slim currently requires manipulation and editing of GO by an expert (or community) familiar with both the ontology and the biological context. Decisions about which terms to include are necessarily subjective, and the creation process itself and subsequent curation are time-consuming and largely manual. Results Here we present an objective framework for generating customised ontology slims for specific annotated datasets, exploiting information latent in the structure of the ontology graph and in the annotation data. This framework combines ontology engineering approaches, and a data-driven algorithm that draws on graph and information theory. We illustrate this method by application to GO, generating GO slims at different information thresholds, characterising their depth of semantics and demonstrating the resulting gains in statistical power. Conclusions Our GO slim creation pipeline is available for use in conjunction with any GO-annotated dataset, and creates dataset-specific, objectively defined slims. This method is fast and scalable for application to other biomedical ontologies.

Motivation: Protein sequences are often composed of regions that have distinct evolutionary histories as a consequence of domain shuffling, recombination or gene conversion. New approaches are required to discover, visualize and analyze these sequence regions and thus enable a better understanding of protein evolution. Results: Here, we have developed an alignment-free and visual approach to analyze sequence relationships. We use the number of shared n-grams between sequences as a measure of sequence similarity and rearrange the resulting affinity matrix applying a spectral technique. Heat maps of the affinity matrix are employed to identify and visualize clusters of related sequences or outliers, while n-gram-based dot plots and conservation profiles allow detailed analysis of similarities among selected sequences. Using this approach, we have identified signatures of domain shuffling in an otherwise poorly characterized family, and homology clusters in another. We conclude that this approach may be generally useful as a framework to analyze related, but highly divergent protein sequences. It is particularly useful as a fast method to study sequence relationships prior to much more time-consuming multiple sequence alignment and phylogenetic analysis. Availability: A software implementation (MOSAIC) of the framework described here can be downloaded from http://bioinformatics.org.au/mosaic/ Contact: m.ragan@uq.edu.au Supplementary information: Supplementary data are available at Bioinformatics online.

The application of the method of matrix representation (MRP) addresses several problems associated with reconstructing phylogenetic relationships among composite organisms such as the eukaryotes. By allowing multiple molecular-sequence and non-molecular data sets to be combined into a single, relatively compact matrix for joint analysis, MRP addresses problems associated with the depth, diversification and specialization of eukaryotic lineages. Characters whose distributions may have been affected by endosymbiotic lateral transfer can be identified by compatibility analysis of MRP-generated hybrid matrices. In conjunction with variant-filtering techniques, MRP can be used to map patterns of lateral transfer in a given phylogenetic tree. These applications are illustrated with molecular-sequence and non-molecular data sets for eukaryotes.

Summary: Genome evolution underpins all of biology, yet its principles can be difficult to communicate to the non-specialist. To facilitate broader understanding of genome evolution, we have designed an interactive 3D environment that enables visualization of diverse genome evolution processes. The system can intuitively and interactively animate mutation histories involving genome rearrangement, point mutation, recombination, insertion and deletion. Multiple organisms related by a phylogeny can be visualized simultaneously. As methods to infer evolutionary histories of genomes become increasingly complex, visualization of the evolutionary process will not only be useful for communication, but will also serve as an exploratory tool for discovering new patterns of genome evolution. Availability: The software is licensed under the GNU GPL and available for download from http://seevolution.org. Contact: aarondarling@ucdavis.edu

Background Many biological processes are mediated by dynamic interactions between and among proteins. In order to interact, two proteins must co-occur spatially and temporally. As protein-protein interactions (PPIs) and subcellular location (SCL) are discovered via separate empirical approaches, PPI and SCL annotations are independent and might complement each other in helping us to understand the role of individual proteins in cellular networks. We expect reliable PPI annotations to show that proteins interacting in vivo are co-located in the same cellular compartment. Our goal here is to evaluate the potential of using PPI annotation in determining SCL of proteins in human, mouse, fly and yeast, and to identify and quantify the factors that contribute to this complementarity. Results Using publicly available data, we evaluate the hypothesis that interacting proteins must be co-located within the same subcellular compartment. Based on a large, manually curated PPI dataset, we demonstrate that a substantial proportion of interacting proteins are in fact co-located. We develop an approach to predict the SCL of a protein based on the SCL of its interaction partners, given sufficient confidence in the interaction itself. The frequency of false positive PPIs can be reduced by use of six lines of supporting evidence, three based on type of recorded evidence (empirical approach, multiplicity of databases, and multiplicity of literature citations) and three based on type of biological evidence (inferred biological process, domain-domain interactions, and orthology relationships), with biological evidence more-effective than recorded evidence. Our approach performs better than four existing prediction methods in identifying the SCL of membrane proteins, and as well as or better for soluble proteins. Conclusion Understanding cellular systems requires knowledge of the SCL of interacting proteins. We show how PPI data can be used more effectively to yield reliable SCL predictions for both soluble and membrane proteins. Scope exists for further improvement in our understanding of cellular function through consideration of the biological context of molecular interactions.

Background: In prokaryotes and some eukaryotes, genetic material can be transferred laterally among unrelated lineages and recombined into new host genomes, providing metabolic and physiological novelty. Although the process is usually framed in terms of gene sharing (e.g. lateral gene transfer, LGT), there is little reason to imagine that the units of transfer and recombination correspond to entire, intact genes. Proteins often consist of one or more spatially compact structural regions (domains) which may fold autonomously and which, singly or in combination, confer the protein's specific functions. As LGT is frequent in strongly selective environments and natural selection is based on function, we hypothesized that domains might also serve as modules of genetic transfer, i.e. that regions of DNA that are transferred and recombined between lineages might encode intact structural domains of proteins. Methodology/Principal Findings: We selected 1,462 orthologous gene sets representing 144 prokaryotic genomes, and applied a rigorous two-stage approach to identify recombination breakpoints within these sequences. Recombination breakpoints are very significantly over-represented in gene sets within which protein domain-encoding regions have been annotated. Within these gene sets, breakpoints significantly avoid the domain-encoding regions (domons), except where these regions constitute most of the sequence length. Recombination breakpoints that fall within longer domons are distributed uniformly at random, but those that fall within shorter domons may show a slight tendency to avoid the domon midpoint. As we find no evidence for differential selection against nucleotide substitutions following the recombination event, any bias against disruption of domains must be a consequence of the recombination event per se. Conclusions/Significance: This is the first systematic study relating the units of LGT to structural features at the protein level. Many genes have been interrupted by recombination following inter-lineage genetic transfer, during which the regions within these genes that encode protein domains have not been preferentially preserved intact. Protein domains are units of function, but domons are not modules of transfer and recombination. Our results demonstrate that LGT can remodel even the most functionally conservative modules within genomes. Copyright: © 2009 Chan et al.

Motivation: The classification of proteins into homologous groups (families) allows their structure and function to be analysed and compared in an evolutionary context. The modular nature of eukaryotic proteins presents a considerable challenge to the delineation of families, as different local regions within a single protein may share common ancestry with distinct, even mutually exclusive, sets of homologs, thereby creating an intricate web of homologous relationships if full-length sequences are taken as the unit of evolution. We attempt to disentangle this web by developing a fully automated pipeline to delineate protein subsequences that represent sensible units for homology inference, and clustering them into putatively homologous families using the Markov clustering algorithm. Results: Using six eukaryotic proteomes as input, we clustered 162 349 protein sequences into 19 697–77 415 subsequence families depending on granularity of clustering. We validated these Markov clusters of homologous subsequences (MACHOS) against the manually curated Pfam domain families, using a quality measure to assess overlap. Our subsequence families correspond well to known domain families and achieve higher quality scores than do groups generated by fully automated domain family classification methods. We illustrate our approach by analysis of a group of proteins that contains the glutamyl/glutaminyl-tRNA synthetase domain, and conclude that our method can produce high-coverage decomposition of protein sequence space into precise homologous families in a way that takes the modularity of eukaryotic proteins into account. This approach allows for a fine-scale examination of evolutionary histories of proteins encoded in eukaryotic genomes. Contact: m.ragan@imb.uq.edu.au Supplementary information: Supplementary data are available at Bioinformatics online. MACHOS for the six proteomes are available as FASTA-formatted files: http://research1t.imb.uq.edu.au/ragan/machos

Nucleotide sequences of microbial genomes provide evidence that genes have been shared among organisms, a phenomenon known as lateral genetic transfer (LGT). Hypotheses about the importance of LGT in the evolution and diversification of microbes can be tested by analyzing the extensive quantities of sequence data now available. Some analysis methods identify genes with sequence features that differ from those of the surrounding genome, whereas other methods are based on inference and comparison of phylogenetic trees. A large-scale search for LGT in 144 genomes using phylogenetic methods has revealed that although parent-to-offspring ("vertical") inheritance has been the dominant mode of gene transmission, LGT has nonetheless been frequent, especially among organisms that are closely related or share the same habitat. This chapter outlines how bioinformatic and phylogenetic analyses can be built into a workflow to identify LGT among microbial genomes.

The identification of homologous DNA is a fundamental buildingblock of comparative genomic and molecular evolution studies. To date, pairwiselocal sequence alignment methods have been the prevailing technique to identifyhomologous nucleotides. However, existing methods that identify and align allhomologous nucleotides in one or more genomes have suffered poor scalabilityand limited accuracy.We propose a novel method that couples a gapped extensionheuristic with a previously described efficient filtration method for local multiplealignment. During gapped extension, we use the MUSCLE implementation ofprogressive multiple alignment with iterative refinement. The resulting gappedextensions potentially contain alignments of unrelated sequence. We detectand remove such undesirable alignments using a hidden Markov model topredict the posterior probability of homology. The HMM emission frequenciesfor nucleotide substitutions can be derived from any strand/species-symmetric nucleotide substitution matrix, and we have developed a method to adapt anarbitrary substitution matrix (i.e. HOXD) to organisms with different G+Ccontent. We evaluate the performance of our method and previous approacheson a hybrid dataset of real genomic DNA with simulated interspersed repeats.Our method outperforms existing methods in terms of sensitivity, positivepredictive value, and localizing boundaries of homology. The described methodshave been implemented in the free, open-source procrastAligner software,available from: http://alggen.lsi.upc.es/recerca/align/procrastination.

Australia enjoys a high international reputation for research in experimental genetics, molecular and cell biology, animal and plant sciences, biotechnology, medicine, biodiversity, and ecological modelling. Computational research is broadly established in these domain areas and others relevant to bioscience. Combined with strong traditions in mathematics and statistics, and national and state investment in computational infrastructure, computational biology Down Under is simply too rich and diverse to be done justice in these few pages. Here we (see Box 1 Authors' Biographies) focus more narrowly, attempting to provide a snapshot of bioinformatic and computational genome-scale biology in Australia circa 2008.

Systems biology is based on computational modelling and simulation of large networks of interacting components. Models may be intended to capture processes, mechanisms, components and interactions at different levels of fidelity. Input data are often large and geographically disperse, and may require the computation to be moved to the data, not vice versa. In addition, complex system-level problems require collaboration across institutions and disciplines. Grid computing can offer robust, scaleable solutions for distributed data, compute and expertise. We illustrate some of the range of computational and data requirements in systems biology with three case studies: one requiring large computation but small data (orthologue mapping in comparative genomics), a second involving complex terabyte data (the Visible Cell project) and a third that is both computationally and data-intensive (simulations at multiple temporal and spatial scales). Authentication, authorisation and audit systems are currently not well scalable and may present bottlenecks for distributed collaboration particularly where outcomes may be commercialised. Challenges remain in providing lightweight standards to facilitate the penetration of robust, scalable grid-type computing into diverse user communities to meet the evolving demands of systems biology.

Background Genetic recombination can produce heterogeneous phylogenetic histories within a set of homologous genes. These recombination events can be obscured by subsequent residue substitutions, which consequently complicate their detection. While there are many algorithms for the identification of recombination events, little is known about the effects of subsequent substitutions on the accuracy of available recombination-detection approaches. Results We assessed the effect of subsequent substitutions on the detection of simulated recombination events within sets of four nucleotide sequences under a homogeneous evolutionary model. The amount of subsequent substitutions per site, prior evolutionary history of the sequences, and reciprocality or non-reciprocality of the recombination event all affected the accuracy of the recombination-detecting programs examined. Bayesian phylogenetic-based approaches showed high accuracy in detecting evidence of recombination event and in identifying recombination breakpoints. These approaches were less sensitive to parameter settings than other methods we tested, making them easier to apply to various data sets in a consistent manner. Conclusion Post-recombination substitutions tend to diminish the predictive accuracy of recombination-detecting programs. The best method for detecting recombined regions is not necessarily the most accurate in identifying recombination breakpoints. For difficult detection problems involving highly divergent sequences or large data sets, different types of approach can be run in succession to increase efficiency, and can potentially yield better predictive accuracy than any single method used in isolation.

Background Bayesian phylogenetic inference holds promise as an alternative to maximum likelihood, particularly for large molecular-sequence data sets. We have investigated the performance of Bayesian inference with empirical and simulated protein-sequence data under conditions of relative branch-length differences and model violation. Results With empirical protein-sequence data, Bayesian posterior probabilities provide more-generous estimates of subtree reliability than does the nonparametric bootstrap combined with maximum likelihood inference, reaching 100% posterior probability at bootstrap proportions around 80%. With simulated 7-taxon protein-sequence datasets, Bayesian posterior probabilities are somewhat more generous than bootstrap proportions, but do not saturate. Compared with likelihood, Bayesian phylogenetic inference can be as or more robust to relative branch-length differences for datasets of this size, particularly when among-sites rate variation is modeled using a gamma distribution. When the (known) correct model was used to infer trees, Bayesian inference recovered the (known) correct tree in 100% of instances in which one or two branches were up to 20-fold longer than the others. At ratios more extreme than 20-fold, topological accuracy of reconstruction degraded only slowly when only one branch was of relatively greater length, but more rapidly when there were two such branches. Under an incorrect model of sequence change, inaccurate trees were sometimes observed at less extreme branch-length ratios, and (particularly for trees with single long branches) such trees tended to be more inaccurate. The effect of model violation on accuracy of reconstruction for trees with two long branches was more variable, but gamma-corrected Bayesian inference nonetheless yielded more-accurate trees than did either maximum likelihood or uncorrected Bayesian inference across the range of conditions we examined. Assuming an exponential Bayesian prior on branch lengths did not improve, and under certain extreme conditions significantly diminished, performance. The two topology-comparison metrics we employed, edit distance and Robinson-Foulds symmetric distance, yielded different but highly complementary measures of performance. Conclusions Our results demonstrate that Bayesian inference can be relatively robust against biologically reasonable levels of relative branch-length differences and model violation, and thus may provide a promising alternative to maximum likelihood for inference of phylogenetic trees from protein-sequence data.

Background Grouping proteins into sequence-based clusters is a fundamental step in many bioinformatic analyses (e.g., homology-based prediction of structure or function). Standard clustering methods such as single-linkage clustering capture a history of cluster topologies as a function of threshold, but in practice their usefulness is limited because unrelated sequences join clusters before biologically meaningful families are fully constituted, e.g. as the result of matches to so-called promiscuous domains. Use of the Markov Cluster algorithm avoids this non-specificity, but does not preserve topological or threshold information about protein families. Results We describe a hybrid approach to sequence-based clustering of proteins that combines the advantages of standard and Markov clustering. We have implemented this hybrid approach over a relational database environment, and describe its application to clustering a large subset of PDB, and to 328577 proteins from 114 fully sequenced microbial genomes. To demonstrate utility with difficult problems, we show that hybrid clustering allows us to constitute the paralogous family of ATP synthase F1 rotary motor subunits into a single, biologically interpretable hierarchical grouping that was not accessible using either single-linkage or Markov clustering alone. We describe validation of this method by hybrid clustering of PDB and mapping SCOP families and domains onto the resulting clusters. Conclusion Hybrid (Markov followed by single-linkage) clustering combines the advantages of the Markov Cluster algorithm (avoidance of non-specific clusters resulting from matches to promiscuous domains) and single-linkage clustering (preservation of topological information as a function of threshold). Within the individual Markov clusters, single-linkage clustering is a more-precise instrument, discerning sub-clusters of biological relevance. Our hybrid approach thus provides a computationally efficient approach to the automated recognition of protein families for phylogenomic analysis.