BACKGROUND: Ab initio prediction of splice sites is an essential step in eukaryotic genome annotation. Recent predictors have exploited Deep Learning algorithms and reliable gene structures from model organisms. However, Deep Learning methods for non-model organisms are lacking. RESULTS: We developed Spliceator to predict splice sites in a wide range of species, including model and non-model organisms. Spliceator uses a convolutional neural network and is trained on carefully validated data from over 100 organisms. We show that Spliceator achieves consistently high accuracy (89-92%) compared to existing methods on independent benchmarks from human, fish, fly, worm, plant and protist organisms. CONCLUSIONS: Spliceator is a new Deep Learning method trained on high-quality data, which can be used to predict splice sites in diverse organisms, ranging from human to protists, with consistently high accuracy.

BACKGROUND: Recent advances in sequencing technologies have led to an explosion in the number of genomes available, but accurate genome annotation remains a major challenge. The prediction of protein-coding genes in eukaryotic genomes is especially problematic, due to their complex exon-intron structures. Even the best eukaryotic gene prediction algorithms can make serious errors that will significantly affect subsequent analyses. RESULTS: We first investigated the prevalence of gene prediction errors in a large set of 176,478 proteins from ten primate proteomes available in public databases. Using the well-studied human proteins as a reference, a total of 82,305 potential errors were detected, including 44,001 deletions, 27,289 insertions and 11,015 mismatched segments where part of the correct protein sequence is replaced with an alternative erroneous sequence. We then focused on the mismatched sequence errors that cause particular problems for downstream applications. A detailed characterization allowed us to identify the potential causes for the gene misprediction in approximately half (5446) of these cases. As a proof-of-concept, we also developed a simple method which allowed us to propose improved sequences for 603 primate proteins. CONCLUSIONS: Gene prediction errors in primate proteomes affect up to 50% of the sequences. Major causes of errors include undetermined genome regions, genome sequencing or assembly issues, and limitations in the models used to represent gene exon-intron structures. Nevertheless, existing genome sequences can still be exploited to improve protein sequence quality. Perspectives of the work include the characterization of other types of gene prediction errors, as well as the development of a more comprehensive algorithm for protein sequence error correction.

Summary: Comparative studies of protein sequences are widely used in evolutionary and comparative genomics studies, but there is a lack of efficient tools to identify conserved regions ab initio within a protein multiple alignment. PROBE provides a fully automatic analysis of protein family conservation, to identify conserved regions, or 'blocks', that may correspond to structural/functional domains or motifs. Conserved blocks are identified at two different levels: (i) family level blocks indicate sites that are probably of central importance to the protein's structure or function, and (ii) sub-family level blocks highlight regions that may signify functional specialization, such as binding partners, etc. All conserved blocks are mapped onto a phylogenetic tree and can also be visualized in the context of the multiple sequence alignment. PROBE thus facilitates in-depth studies of sequence-structure-function-evolution relationships, and opens the way to block-level phylogenetic profiling. Availability and implementation: Freely available on the web at http://www.lbgi.fr/∼julie/probe/web.

BACKGROUND: A standard procedure in many areas of bioinformatics is to use a multiple sequence alignment (MSA) as the basis for various types of homology-based inference. Applications include 3D structure modelling, protein functional annotation, prediction of molecular interactions, etc. These applications, however sophisticated, are generally highly sensitive to the alignment used, and neglecting non-homologous or uncertain regions in the alignment can lead to significant bias in the subsequent inferences. RESULTS: Here, we present a new method, LEON-BIS, which uses a robust Bayesian framework to estimate the homologous relations between sequences in a protein multiple alignment. Sequences are clustered into sub-families and relations are predicted at different levels, including 'core blocks', 'regions' and full-length proteins. The accuracy and reliability of the predictions are demonstrated in large-scale comparisons using well annotated alignment databases, where the homologous sequence segments are detected with very high sensitivity and specificity. CONCLUSIONS: LEON-BIS uses robust Bayesian statistics to distinguish the portions of multiple sequence alignments that are conserved either across the whole family or within subfamilies. LEON-BIS should thus be useful for automatic, high-throughput genome annotations, 2D/3D structure predictions, protein-protein interaction predictions etc.

SUMMARY: We previously developed OrthoInspector, a package incorporating an original algorithm for the detection of orthology and inparalogy relations between different species. We have added new functionalities to the package. While its original algorithm was not modified, performing similar orthology predictions, we facilitated the prediction of very large databases (thousands of proteomes), refurbished its graphical interface, added new visualization tools for comparative genomics/protein family analysis and facilitated its deployment in a network environment. Finally, we have released three online databases of precomputed orthology relationships. AVAILABILITY: Package and databases are freely available at http://lbgi.fr/orthoinspector with all major browsers supported. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

MOTIVATION: The prediction of protein coding genes is a major challenge that depends on the quality of genome sequencing, the accuracy of the model used to elucidate the exonic structure of the genes and the complexity of the gene splicing process leading to different protein variants. As a consequence, today's protein databases contain a huge amount of inconsistency, due to both natural variants and sequence prediction errors. RESULTS: We have developed a new method, called SIBIS, to detect such inconsistencies based on the evolutionary information in multiple sequence alignments. A Bayesian framework, combined with Dirichlet mixture models, is used to estimate the probability of observing specific amino acids and to detect inconsistent or erroneous sequence segments. We evaluated the performance of SIBIS on a reference set of protein sequences with experimentally validated errors and showed that the sensitivity is significantly higher than previous methods, with only a small loss of specificity. We also assessed a large set of human sequences from the UniProt database and found evidence of inconsistency in 48% of the previously uncharacterized sequences. We conclude that the integration of quality control methods like SIBIS in automatic analysis pipelines will be critical for the robust inference of structural, functional and phylogenetic information from these sequences. AVAILABILITY AND IMPLEMENTATION: Source code, implemented in C on a linux system, and the datasets of protein sequences are freely available for download at http://www.lbgi.fr/∼julie/SIBIS.

BACKGROUND: Small insertion and deletion polymorphisms (Indels) are the second most common mutations in the human genome, after Single Nucleotide Polymorphisms (SNPs). Recent studies have shown that they have significant influence on genetic variation by altering human traits and can cause multiple human diseases. In particular, many Indels that occur in protein coding regions are known to impact the structure or function of the protein. A major challenge is to predict the effects of these Indels and to distinguish between deleterious and neutral variants. When an Indel occurs within a coding region, it can be either frameshifting (FS) or non-frameshifting (NFS). FS-Indels either modify the complete C-terminal region of the protein or result in premature termination of translation. NFS-Indels insert/delete multiples of three nucleotides leading to the insertion/deletion of one or more amino acids. RESULTS: In order to study the relationships between NFS-Indels and Mendelian diseases, we characterized NFS-Indels according to numerous structural, functional and evolutionary parameters. We then used these parameters to identify specific characteristics of disease-causing and neutral NFS-Indels. Finally, we developed a new machine learning approach, KD4i, that can be used to predict the phenotypic effects of NFS-Indels. CONCLUSIONS: We demonstrate in a large-scale evaluation that the accuracy of KD4i is comparable to existing state-of-the-art methods. However, a major advantage of our approach is that we also provide the reasons for the predictions, in the form of a set of rules. The rules are interpretable by non-expert humans and they thus represent new knowledge about the relationships between the genotype and phenotypes of NFS-Indels and the causative molecular perturbations that result in the disease.

SUMMARY: We present PARSEC (PAtteRn Search and Contextualization), a new open source platform for guided discovery, allowing localization and biological characterization of short genomic sites in entire eukaryotic genomes. PARSEC can search for a sequence or a degenerated pattern. The retrieved set of genomic sites can be characterized in terms of (i) conservation in model organisms, (ii) genomic context (proximity to genes) and (iii) function of neighboring genes. These modules allow the user to explore, visualize, filter and extract biological knowledge from a set of short genomic regions such as transcription factor binding sites. AVAILABILITY: Web site implemented in Java, JavaScript and C++, with all major browsers supported. Freely available at lbgi.fr/parsec. Source code is freely available at sourceforge.net/projects/genomicparsec.

BACKGROUND: The accurate determination of orthology and inparalogy relationships is essential for comparative sequence analysis, functional gene annotation and evolutionary studies. Various methods have been developed based on either simple blast all-versus-all pairwise comparisons and/or time-consuming phylogenetic tree analyses. RESULTS: We have developed OrthoInspector, a new software system incorporating an original algorithm for the rapid detection of orthology and inparalogy relations between different species. In comparisons with existing methods, OrthoInspector improves detection sensitivity, with a minimal loss of specificity. In addition, several visualization tools have been developed to facilitate in-depth studies based on these predictions. The software has been used to study the orthology/in-paralogy relationships for a large set of 940,855 protein sequences from 59 different eukaryotic species. CONCLUSION: OrthoInspector is a new software system for orthology/paralogy analysis. It is made available as an independent software suite that can be downloaded and installed for local use. Command line querying facilitates the integration of the software in high throughput processing pipelines and a graphical interface provides easy, intuitive access to results for the non-expert.

BACKGROUND: The Gene Ontology (GO) is a well known controlled vocabulary describing the biological process, molecular function and cellular component aspects of gene annotation. It has become a widely used knowledge source in bioinformatics for annotating genes and measuring their semantic similarity. These measures generally involve the GO graph structure, the information content of GO aspects, or a combination of both. However, only a few of the semantic similarity measures described so far can handle GO annotations differently according to their origin (i.e. their evidence codes). RESULTS: We present here a new semantic similarity measure called IntelliGO which integrates several complementary properties in a novel vector space model. The coefficients associated with each GO term that annotates a given gene or protein include its information content as well as a customized value for each type of GO evidence code. The generalized cosine similarity measure, used for calculating the dot product between two vectors, has been rigorously adapted to the context of the GO graph. The IntelliGO similarity measure is tested on two benchmark datasets consisting of KEGG pathways and Pfam domains grouped as clans, considering the GO biological process and molecular function terms, respectively, for a total of 683 yeast and human genes and involving more than 67,900 pair-wise comparisons. The ability of the IntelliGO similarity measure to express the biological cohesion of sets of genes compares favourably to four existing similarity measures. For inter-set comparison, it consistently discriminates between distinct sets of genes. Furthermore, the IntelliGO similarity measure allows the influence of weights assigned to evidence codes to be checked. Finally, the results obtained with a complementary reference technique give intermediate but correct correlation values with the sequence similarity, Pfam, and Enzyme classifications when compared to previously published measures. CONCLUSIONS: The IntelliGO similarity measure provides a customizable and comprehensive method for quantifying gene similarity based on GO annotations. It also displays a robust set-discriminating power which suggests it will be useful for functional clustering. AVAILABILITY: An on-line version of the IntelliGO similarity measure is available at: http://bioinfo.loria.fr/Members/benabdsi/intelligo_project/

The traditional approach to bioinformatics analyses relies on independent task-specific services and applications, using different input and output formats, often idiosyncratic, and frequently not designed to inter-operate. In general, such analyses were performed by experts who manually verified the results obtained at each step in the process. Today, the amount of bioinformatics information continuously being produced means that handling the various applications used to study this information presents a major data management and analysis challenge to researchers. It is now impossible to manually analyse all this information and new approaches are needed that are capable of processing the large-scale heterogeneous data in order to extract the pertinent information. We review the recent use of integrated expert systems aimed at providing more efficient knowledge extraction for bioinformatics research. A general methodology for building knowledge-based expert systems is described, focusing on the unstructured information management architecture, UIMA, which provides facilities for both data and process management. A case study involving a multiple alignment expert system prototype called AlexSys is also presented.

BACKGROUND: Phylogenetic profiling encompasses an important set of methodologies for in silico high throughput inference of functional relationships between genes. The simplest profiles represent the distribution of gene presence-absence in a set of species as a sequence of 0's and 1's, and it is assumed that functionally related genes will have more similar profiles. The methodology has been successfully used in numerous studies of prokaryotic genomes, although its application in eukaryotes appears problematic, with reported low accuracy due to the complex genomic organization within this domain of life. Recently some groups have proposed an alternative approach based on the correlation of homologous gene group sizes, taking into account all potentially informative genetic events leading to a change in group size, regardless of whether they result in a de novo group gain or total gene group loss. RESULTS: We have compared the performance of classical presence-absence and group size based approaches using a large, diverse set of eukaryotic species. In contrast to most previous comparisons in Eukarya, we take into account the species phylogeny. We also compare the approaches using two different group categories, based on orthology and on domain-sharing. Our results confirm a limited overall performance of phylogenetic profiling in eukaryotes. Although group size based approaches initially showed an increase in performance for the domain-sharing based groups, this seems to be an overestimation due to a simplistic negative control dataset and the choice of null hypothesis rejection criteria. CONCLUSION: Presence-absence profiling represents a more accurate classifier of related versus non-related profile pairs, when the profiles under consideration have enough information content. Group size based approaches provide a complementary means of detecting domain or family level co-evolution between groups that may be elusive to presence-absence profiling. Moreover positive correlation between co-evolution scores and functional links imply that these methods could be used to estimate functional distances between gene groups and to cluster them based on their functional relatedness. This study should have important implications for the future development and application of phylogenetic profiling methods, not only in eukaryotic, but also in prokaryotic datasets.

UNLABELLED: With the establishment of high-throughput (HT) screening methods there is an increasing need for automatic analysis methods. Here we present RReportGenerator, a user-friendly portal for automatic routine analysis using the statistical platform R and Bioconductor. RReportGenerator is designed to analyze data using predefined analysis scenarios via a graphical user interface (GUI). A report in pdf format combining text, figures and tables is automatically generated and results may be exported. To demonstrate suitable analysis tasks we provide direct web access to a collection of analysis scenarios for summarizing data from transfected cell arrays (TCA), segmentation of CGH data, and microarray quality control and normalization. AVAILABILITY: RReportGenerator, a user manual and a collection of analysis scenarios are available under a GNU public license on http://www-bio3d-igbmc.u-strasbg.fr/~wraff

BACKGROUND: Linear motifs (LMs) are abundant short regulatory sites used for modulating the functions of many eukaryotic proteins. They play important roles in post-translational modification, cell compartment targeting, docking sites for regulatory complex assembly and protein processing and cleavage. Methods for LM detection are now being developed that are strongly dependent on scores for motif conservation in homologous proteins. However, most LMs are found in natively disordered polypeptide segments that evolve rapidly, unhindered by structural constraints on the sequence. These regions of modular proteins are difficult to align using classical multiple sequence alignment programs that are specifically optimised to align the globular domains. As a consequence, poor motif alignment quality is hindering efforts to detect new LMs. RESULTS: We have developed a new benchmark, as part of the BAliBASE suite, designed to assess the ability of standard multiple alignment methods to detect and align LMs. The reference alignments are organised into different test sets representing real alignment problems and contain examples of experimentally verified functional motifs, extracted from the Eukaryotic Linear Motif (ELM) database. The benchmark has been used to evaluate and compare a number of multiple alignment programs. With distantly related proteins, the worst alignment program correctly aligns 48% of LMs compared to 73% for the best program. However, the performance of all the programs is adversely affected by the introduction of other sequences containing false positive motifs. The ranking of the alignment programs based on LM alignment quality is similar to that observed when considering full-length protein alignments, however little correlation was observed between LM and overall alignment quality for individual alignment test cases. CONCLUSION: We have shown that none of the programs currently available is capable of reliably aligning LMs in distantly related sequences and we have highlighted a number of specific problems. The results of the tests suggest possible ways to improve program accuracy for difficult, divergent sequences.

The amino acid sequence of a protein is the key to understanding its structure and ultimately its function in the cell. This paper addresses the fundamental issue of encoding amino acids in ways that the visualization of protein sequences facilitates the decoding of its information content. We show that a feature-based representation in a three-dimensional (3D) space derived from substitution matrices provides an adequate representation from which the domain content of a protein can be predicted. In addition, we show that each dimension of the feature space can be related to a physical property of the amino acids.

BACKGROUND: The post-genomic era is characterised by a torrent of biological information flooding the public databases. As a direct consequence, similarity searches starting with a single query sequence frequently lead to the identification of hundreds, or even thousands of potential homologues. The huge volume of data renders the subsequent structural, functional and evolutionary analyses very difficult. It is therefore essential to develop new strategies for efficient sampling of this large sequence space, in order to reduce the number of sequences to be processed. At the same time, it is important to retain the most pertinent sequences for structural and functional studies. RESULTS: An exhaustive analysis on a large scale test set (284 protein families) was performed to compare the efficiency of four different sampling methods aimed at selecting the most pertinent sequences. These four methods sample the proteins detected by BlastP searches and can be divided into two categories: two customisable methods where the user defines either the maximal number or the percentage of sequences to be selected; two automatic methods in which the number of sequences selected is determined by the program. We focused our analysis on the potential information content of the sampled sets of sequences using multiple alignment of complete sequences as the main validation tool. The study considered two criteria: the total number of sequences in BlastP and their associated E-values. The subsequent analyses investigated the influence of the sampling methods on the E-value distributions, the sequence coverage, the final multiple alignment quality and the active site characterisation at various residue conservation thresholds as a function of these criteria. CONCLUSION: The comparative analysis of the four sampling methods allows us to propose a suitable sampling strategy that significantly reduces the number of homologous sequences required for alignment, while at the same time maintaining the relevant information concerning the active site residues.

Abstract Summary: MAGOS is a web server allowing automated protein modelling coupled to the creation of a hierarchical and annotated multiple alignment of complete sequences. MAGOS is designed for an interactive approach of structural information within the framework of the evolutionary relevance of mined and predicted sequence information. Availability: The web server is freely available at Supplementary information: The website supplies detailed explanations and illustrations of processes and results at Contact: [email protected]

BACKGROUND: In the post-genomic era, systems-level studies are being performed that seek to explain complex biological systems by integrating diverse resources from fields such as genomics, proteomics or transcriptomics. New information management systems are now needed for the collection, validation and analysis of the vast amount of heterogeneous data available. Multiple alignments of complete sequences provide an ideal environment for the integration of this information in the context of the protein family. RESULTS: MACSIMS is a multiple alignment-based information management program that combines the advantages of both knowledge-based and ab initio sequence analysis methods. Structural and functional information is retrieved automatically from the public databases. In the multiple alignment, homologous regions are identified and the retrieved data is evaluated and propagated from known to unknown sequences with these reliable regions. In a large-scale evaluation, the specificity of the propagated sequence features is estimated to be >99%, i.e. very few false positive predictions are made. MACSIMS is then used to characterise mutations in a test set of 100 proteins that are known to be involved in human genetic diseases. The number of sequence features associated with these proteins was increased by 60%, compared to the features available in the public databases. An XML format output file allows automatic parsing of the MACSIM results, while a graphical display using the JalView program allows manual analysis. CONCLUSION: MACSIMS is a new information management system that incorporates detailed analyses of protein families at the structural, functional and evolutionary levels. MACSIMS thus provides a unique environment that facilitates knowledge extraction and the presentation of the most pertinent information to the biologist. A web server and the source code are available at http://bips.u-strasbg.fr/MACSIMS/.

UNLABELLED: GOAnno is a web tool that automatically annotates proteins according to the Gene Ontology (GO) using evolutionary information available in hierarchized multiple alignments. GO terms present in the aligned functional subfamily can be cross-validated and propagated to obtain highly reliable predicted GO annotation based on the GOAnno algorithm. AVAILABILITY: The web tool and a reduced version for local installation are freely available at http://igbmc.u-strasbg.fr/GOAnno/GOAnno.html SUPPLEMENTARY INFORMATION: The website supplies a detailed explanation and illustration of the algorithm at http://igbmc.u-strasbg.fr/GOAnno/GOAnnoHelp.html.

MOTIVATION: Recent advances in gene sequencing have provided complete sequence information for a number of genomes and as a result the amount of data in the sequence databases is growing at an exponential rate. We introduce here a new program, DbW, to automate the update of a functional family-specific multiple alignment that tries to include relevant sequences. The program is based on the use of different sources of information: sequences and annotations in databases. RESULTS: The advantages of DbW are demonstrated using the 20 families of aminoacyl-tRNA synthetases, where DbW detects a maximum of homologous sequences in the Swiss-Prot and SPTREMBL databases. The global specificity of DbW in this test is 98.4% (1.6% of the sequences included in the alignment did not belong to the family according to their function), and the global sensitivity of DbW is estimated to be 95.2%. Thus, DbW provides a reliable basis for the many applications that rely on accurate multiple alignments, e.g. functional residue identification, 2D/3D structure prediction or homology modeling. AVAILABILITY: The DbW software is available for download at ftp://ftp-igbmc.u-strasbg.fr/pub/DbW/DbW.tar and online at http://titus.u-strasbg.fr/DbW CONTACT: [email protected].

MOTIVATION: Most multiple sequence alignment programs use heuristics that sometimes introduce errors into the alignment. The most commonly used methods to correct these errors use iterative techniques to maximize an objective function. We present here an alternative, knowledge-based approach that combines a number of recently developed methods into a two-step refinement process. The alignment is divided horizontally and vertically to form a 'lattice' in which well aligned regions can be differentiated. Alignment correction is then restricted to the less reliable regions, leading to a more reliable and efficient refinement strategy. RESULTS: The accuracy and reliability of RASCAL is demonstrated using: (i) alignments from the BAliBASE benchmark database, where significant improvements were often observed, with no deterioration of the existing high-quality regions, (ii) a large scale study involving 946 alignments from the ProDom protein domain database, where alignment quality was increased in 68% of the cases; and (iii) an automatic pipeline to obtain a high-quality alignment of 695 full-length nuclear receptor proteins, which took 11 min on a DEC Alpha 6100 computer AVAILABILITY: RASCAL is available at ftp://ftp-igbmc.u-strasbg.fr/pub/RASCAL. SUPPLEMENTARY INFORMATION: http://bioinfo-igbmc.u-strasbourg.fr/BioInfo/RASCAL/paper/rascal_supp.html

MOTIVATION: Blast programs are very efficient in finding relatively strong similarities but some very distantly related sequences are given a very high Expect value and are ranked very low in Blast results. We have developed Ballast, a program to predict local maximum segments (LMSs-i.e. sequence segments conserved relatively to their flanking regions) from a single Blast database search and to highlight these divergent homologues. The TBlastN database searches can also be processed with the help of information from a joint BlastP search. RESULTS: We have applied the Ballast algorithm to BlastP searches performed with sequences belonging to well described dispersed families (aminoacyl-tRNA synthetases; helicases) against the SwissProt 38 database. We show that Ballast is able to build an appropriate conservation profile and that LMSs are predicted that are consistent with the signatures and motifs described in the literature. Furthermore, by comparing the Blast, PsiBlast and Ballast results obtained on a well defined database of structurally related sequences, we show that the LMSs provide a scoring scheme that can concentrate on top ranking distant homologues better than Blast. Using the graphical user interface available on the Web, specific LMSs may be selected to detect divergent homologues sharing the corresponding properties with the query sequence without requiring any additional database search.

SUMMARY: BAliBASE is a database of manually refined multiple sequence alignments categorized by core blocks of conservation sequence length, similarity, and the presence of insertions and N/C-terminal extensions. AVAILABILITY: From http://www-igbmc. u-strasbg.fr/BioInfo/BAliBASE/index.html