Search

Abstract Background With the growing amount of biomedical data available in public databases it has become increasingly important to annotate data in a consistent way in order to allow easy access to this rich source of information. Annotating the data using controlled vocabulary terms and ontologies makes it much easier to compare and analyze data from different sources. However, finding the correct controlled vocabulary terms can sometimes be a difficult task for the end user annotating these data. Results In order to facilitate the location of the correct term in the correct controlled vocabulary or ontology, the Ontology Lookup Service was created. However, using the Ontology Lookup Service as a web service is not always feasible, especially for researchers without bioinformatics support. We have therefore created a Java front end to the Ontology Lookup Service, called the OLS Dialog, which can be plugged into any application requiring the annotation of data using controlled vocabulary terms, making it possible to find and use controlled vocabulary terms without requiring any additional knowledge about web services or ontology formats. Conclusions As a user-friendly open source front end to the Ontology Lookup Service, the OLS Dialog makes it straightforward to include controlled vocabulary support in third-party tools, which ultimately makes the data even more valuable to the biomedical community.

Abstract Background Mass spectrometric analysis of peptides is an essential part of protein identification and characterization, the latter meaning the identification of modifications and amino acid substitutions. There are two main approaches for characterization: (i) using a predefined set of possible modifications and substitutions or (ii) performing a blind search. The first option is straightforward, but can not detect modifications or substitutions outside the predefined set. A blind search does not have this limitation, and therefore has the potential of detecting both known and unknown modifications and substitutions. Combining the peptide mass fingerprints from two proteases result in overlapping sequence coverage of the protein, thereby offering alternative views of the protein and a novel way of indicating post-translational modifications and amino acid substitutions. Results We have developed an algorithm and a software tool, MassShiftFinder, that performs a blind search using peptide mass fingerprints from two proteases with different cleavage specificities. The algorithm is based on equal mass shifts for overlapping peptides from the two proteases used, and can indicate both post-translational modifications and amino acid substitutions. In most cases it is possible to suggest a restricted area within the overlapping peptides where the mass shift can occur. The program is available at http://www.bioinfo.no/software/massShiftFinder. Conclusion Without any prior assumptions on their presence the described algorithm is able to indicate post-translational modifications or amino acid substitutions in MALDI-TOF experiments on identified proteins, and can thereby direct the involved peptides to subsequent TOF-TOF analysis. The algorithm is designed for detailed and low-throughput characterization of single proteins.

Abstract Background Proteomics is the study of the proteome, and is critical to the understanding of cellular processes. Two central and related tasks of proteomics are protein identification and protein characterization. Many small laboratories are interested in the characterization of a small number of proteins, e.g., how posttranslational modifications change under different conditions. Results We have developed a software tool called MassSorter for administrating and analyzing data from peptide mass fingerprinting experiments on proteins with known amino acid sequences. It is meant for small scale mass spectrometry laboratories that are interested in posttranslational modifications of known proteins. Several experiments can be compared simultaneously, and the matched and unmatched peak values are clearly indicated. The hits can be sorted according to m/z values (default) or according to the sequence of the protein. Filters defined by the user can mark autolytic protease peaks and other contaminating peaks (keratins, proteins co-migrating with the protein of interest, etc.). Unmatched peaks can be further analyzed for unexpected modifications by searches against a local version of the UniMod database. They can also be analyzed for unexpected cleavages, a highly useful feature for proteins that undergo maturation by proteolytic cleavage, creating new N- or C-terminals. Additional tools exist for visualization of the results, like sequence coverage, accuracy plots, different types of statistics, 3D models, etc. The program and a tutorial are freely available for academic users at http://www.bioinfo.no/software/massSorter. Conclusion MassSorter has a number of useful features that can promote the analysis and administration of MS-data.

Abstract Background As genomes evolve after speciation, gene content, coding sequence, gene expression, and splicing all diverge with time from ancestors with close relatives. A minimum evolution general method for continuous character analysis in a phylogenetic perspective is presented that allows for reconstruction of ancestral character states and for measuring along branch evolution. Results A software package for reconstruction of continuous character traits, like relative gene expression levels or alternative splice site usage data is presented and is available for download at http://www.rossnes.org/phyrex. This program was applied to a primate gene expression dataset to detect transcription factor binding sites that have undergone substitution, potentially having driven lineage-specific differences in gene expression. Conclusion Systematic analysis of lineage-specific evolution is becoming the cornerstone of comparative genomics. New methods, like phyrex, extend the capabilities of comparative genomics by tracing the evolution of additional biomolecular processes.

Understanding the fragmentation process in MS/MS experiments is vital when trying to validate the results of such experiments, and one way of improving our understanding is to analyze existing data. We here present our findings from an analysis of a large and diverse data set of MS/MS-based peptide identifications, in which each peptide has been identified from multiple spectra, recorded on two commonly used types of electrospray instruments. By analyzing these data we were able to study fragmentation variability on three levels: (i) variation in detection rates and intensities for fragment ions from the same peptide sequence measured multiple times on a single instrument; (ii) consistency of rank-based fragmentation patterns; and (iii) a set of general observations on fragment ion occurrence in MS/MS experiments, regardless of sequence. Our results confirm that substantial variation can be found at all levels, even when high-quality identifications are used and the experimental conditions as well as the peptide sequences are kept constant. Finally, we discuss the observed variability in light of ongoing efforts to create spectral libraries and predictive software for target selection in targeted proteomics., (Copyright © 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.)

With the continuously growing amount of proteomics data being produced, it has become increasingly important to make these data publicly available so that they can be audited, reanalyzed, and reused. More and more journals are also starting to request the deposition of MS data in publicly available repositories for submitted proteomics manuscripts. In this chapter we focus on one of the most commonly used proteomics data repositories, PRIDE (the PRoteomics IDEntifications database, http://www.ebi.ac.uk/pride), and demonstrate how a new graphical user interface tool called PRIDE Converter (http://pride-converter.googlecode.com) greatly simplifies the submission of data to PRIDE.

A thorough understanding of the fragmentation processes in MS/MS can be a powerful tool in assessing the resulting peptide and protein identifications. We here present the freely available, open-source FragmentationAnalyzer tool (http://fragmentation-analyzer.googlecode.com) that makes it straightforward to analyze large MS/MS data sets for specific types of identified peptides, using a common set of peptide properties. This enables the detection of fragmentation pattern nuances related to specific instruments or due to the presence of post-translational modifications.

Background: With the growing amount of biomedical data available in public databases it has become increasingly important to annotate data in a consistent way in order to allow easy access to this rich source of information. Annotating the data using controlled vocabulary terms and ontologies makes it much easier to compare and analyze data from different sources. However, finding the correct controlled vocabulary terms can sometimes be a difficult task for the end user annotating these data., Results: In order to facilitate the location of the correct term in the correct controlled vocabulary or ontology, the Ontology Lookup Service was created. However, using the Ontology Lookup Service as a web service is not always feasible, especially for researchers without bioinformatics support. We have therefore created a Java front end to the Ontology Lookup Service, called the OLS Dialog, which can be plugged into any application requiring the annotation of data using controlled vocabulary terms, making it possible to find and use controlled vocabulary terms without requiring any additional knowledge about web services or ontology formats., Conclusions: As a user-friendly open source front end to the Ontology Lookup Service, the OLS Dialog makes it straightforward to include controlled vocabulary support in third-party tools, which ultimately makes the data even more valuable to the biomedical community.

Protein identification by MS is an important technique in both gel-based and gel-free proteome studies. The Open Mass Spectrometry Search Algorithm (OMSSA) (http://pubchem.ncbi.nlm.nih.gov/omssa) is an open-source search engine that can be used to identify MS/MS spectra acquired in these experiments. We here present a lightweight, open-source Java software library, OMSSA Parser (http://code.google.com/p/omssa-parser), which parses OMSSA omx result files into easy accessible and fully functional object models. In addition, we also provide examples illustrating the usage of our library.

Background: Mass spectrometric analysis of peptides is an essential part of protein identification and characterization, the latter meaning the identification of modifications and amino acid substitutions. There are two main approaches for characterization: (i) using a predefined set of possible modifications and substitutions or (ii) performing a blind search. The first option is straightforward, but can not detect modifications or substitutions outside the predefined set. A blind search does not have this limitation, and therefore has the potential of detecting both known and unknown modifications and substitutions. Combining the peptide mass fingerprints from two proteases result in overlapping sequence coverage of the protein, thereby offering alternative views of the protein and a novel way of indicating post-translational modifications and amino acid substitutions., Results: We have developed an algorithm and a software tool, MassShiftFinder, that performs a blind search using peptide mass fingerprints from two proteases with different cleavage specificities. The algorithm is based on equal mass shifts for overlapping peptides from the two proteases used, and can indicate both post-translational modifications and amino acid substitutions. In most cases it is possible to suggest a restricted area within the overlapping peptides where the mass shift can occur. The program is available at http://www.bioinfo.no/software/massShiftFinder., Conclusion: Without any prior assumptions on their presence the described algorithm is able to indicate post-translational modifications or amino acid substitutions in MALDI-TOF experiments on identified proteins, and can thereby direct the involved peptides to subsequent TOF-TOF analysis. The algorithm is designed for detailed and low-throughput characterization of single proteins.

MassSorter is a software tool that sorts, systemizes, and analyzes data from peptide mass fingerprinting (PMF) experiments on proteins with known amino acid sequences. Several experiments can be simultaneously analyzed for sequence coverage and posttranslational modifications occurring during sample handling, induced chemical modifications, and unexpected cleavages. Experimental m/z values are compared with m/z values from an in silico digestion, taking modifications into account. Filters can be defined by users for marking autolytic protease peaks and other contaminating peaks. MassSorter functions as a database of all the detected peptides. It includes tools for visualization of the results, such as sequence coverage, accuracy plots, statistics, and 3D models.

Mass spectrometric analyses of peptides mainly rely on cleavage of proteins with proteases that have a defined specificity. The specificities of the proteases imply that there is not a random distribution of amino acids in the peptides. The physico-chemical effects of this distribution have been partly analyzed for tryptic peptides, but to a lesser degree for other proteases. Using all human proteins in Swiss-Prot, the relationships between peptide fractional mass, pI and hydrophobicity were investigated. The distribution of the fractional masses and the average regression lines for the fractional masses were similar, but not identical, for the peptides generated by the proteases trypsin, chymotrypsin and gluC, with the steepest regression line for gluC. The fractional mass regression lines for individual proteins showed up to +/-100 ppm in mass difference from the average regression line and the peptides generated showed protease-dependent properties. We here show that the fractional mass and some other properties of the peptides are dependent on the protease used for generating the peptides. With the increasing accuracy of mass spectrometry instruments it is possible to exploit the information embedded in the fractional mass of unknown peaks in peptide mass fingerprint spectra.

High-throughput proteomics experiments typically generate large amounts of peptide fragmentation mass spectra during a single experiment. There is often a substantial amount of redundant fragmentation of the same precursors among these spectra, which is usually considered a nuisance. We here discuss the potential of clustering and merging redundant spectra to turn this redundancy into a useful property of the dataset. To this end, we have created the first general-purpose, freely available open-source software application for clustering and merging MS/MS spectra. The application also introduces a novel approach to calculating the similarity of fragmentation mass spectra that takes into account the increased precision of modern mass spectrometers, and we suggest a simple but effective improvement to single-linkage clustering. The application and the novel algorithms are applied to several real-life proteomic datasets and the results are discussed. An analysis of the influence of the different algorithms available and their parameters is given, as well as a number of important applications of the overall approach.

A novel alignment-free method for computing functional similarity of membrane proteins based on features of hydropathy distribution is presented. The features of hydropathy distribution are used to represent protein families as hydropathy profiles. The profiles statistically summarize the hydropathy distribution of member proteins. The summation is made by using hydropathy features that numerically represent structurally/functionally significant portions of protein sequences. The hydropathy profiles are numerical vectors that are points in a high dimensional 'hydropathy' space. Their similarities are identified by projection of the space onto principal axes. Here, the approach is applied to the secondary transporters. The analysis using the presented approach is validated by the standard classification of the secondary transporters. The presented analysis allows for prediction of function attributes for proteins of uncharacterized families of secondary transporters. The results obtained using the presented analysis may help to characterize unknown function attributes of secondary transporters. They also show that analysis of hydropathy distribution can be used for function prediction of membrane proteins.

In contemporary peptide-centric or non-gel proteome studies, vast amounts of peptide fragmentation data are generated of which only a small part leads to peptide or protein identification. This motivates the development and use of a filtering algorithm that removes spectra that contribute little to protein identification. Removal of unidentifiable spectra reduced both the amount of computational and human time spent on analyzing spectra as well as the chances of obtaining false identifications. Thorough testing on various proteome datasets from different instruments showed that the best suggested machine-learning classifier is, on average, able to recognize half of the unidentified spectra as bad spectra. Further analyses showed that several unidentified spectra classified as good were derived from peptides carrying unanticipated amino acid modifications or contained sequence tags that allowed peptide identification using homology searches. The implementation of the classifiers is available under the GNU General Public License at http://www.bioinfo.no/software/spectrumquality.

High-resolution two-dimensional gel electrophoresis and mass spectrometry has been used to identify the outer membrane (OM) subproteome of the Gram-negative bacterium Methylococcus capsulatus (Bath). Twenty-eight unique polypeptide sequences were identified from protein samples enriched in OMs. Only six of these polypeptides had previously been identified. The predictions from novel bioinformatic methods predicting beta-barrel outer membrane proteins (OMPs) and OM lipoproteins were compared to proteins identified experimentally. BOMP ( http://www.bioinfo.no/tools/bomp ) predicted 43 beta-barrel OMPs (1.45%) from the 2,959 annotated open reading frames. This was a lower percentage than predicted from other Gram-negative proteomes (1.8-3%). More than half of the predicted BOMPs in M. capsulatus were annotated as (conserved) hypothetical proteins with significant similarity to very few sequences in Swiss-Prot or TrEMBL. The experimental data and the computer predictions indicated that the protein composition of the M. capsulatus OM subproteome was different from that of other Gram-negative bacteria studied in a similar manner. A new program, Lipo, was developed that can analyse entire predicted proteomes and give a list of recognised lipoproteins categorised according to their lipo-box similarity to known Gram-negative lipoproteins ( http://www.bioinfo.no/tools/lipo ). This report is the first using a proteomics and bioinformatics approach to identify the OM subproteome of an obligate methanotroph.

Background: Proteomics is the study of the proteome, and is critical to the understanding of cellular processes. Two central and related tasks of proteomics are protein identification and protein characterization. Many small laboratories are interested in the characterization of a small number of proteins, e.g., how posttranslational modifications change under different conditions., Results: We have developed a software tool called MassSorter for administrating and analyzing data from peptide mass fingerprinting experiments on proteins with known amino acid sequences. It is meant for small scale mass spectrometry laboratories that are interested in posttranslational modifications of known proteins. Several experiments can be compared simultaneously, and the matched and unmatched peak values are clearly indicated. The hits can be sorted according to m/z values (default) or according to the sequence of the protein. Filters defined by the user can mark autolytic protease peaks and other contaminating peaks (keratins, proteins co-migrating with the protein of interest, etc.). Unmatched peaks can be further analyzed for unexpected modifications by searches against a local version of the UniMod database. They can also be analyzed for unexpected cleavages, a highly useful feature for proteins that undergo maturation by proteolytic cleavage, creating new N- or C-terminals. Additional tools exist for visualization of the results, like sequence coverage, accuracy plots, different types of statistics, 3D models, etc. The program and a tutorial are freely available for academic users at http://www.bioinfo.no/software/massSorter., Conclusion: MassSorter has a number of useful features that can promote the analysis and administration of MS-data.

Background: As genomes evolve after speciation, gene content, coding sequence, gene expression, and splicing all diverge with time from ancestors with close relatives. A minimum evolution general method for continuous character analysis in a phylogenetic perspective is presented that allows for reconstruction of ancestral character states and for measuring along branch evolution., Results: A software package for reconstruction of continuous character traits, like relative gene expression levels or alternative splice site usage data is presented and is available for download at http://www.rossnes.org/phyrex. This program was applied to a primate gene expression dataset to detect transcription factor binding sites that have undergone substitution, potentially having driven lineage-specific differences in gene expression., Conclusion: Systematic analysis of lineage-specific evolution is becoming the cornerstone of comparative genomics. New methods, like phyrex, extend the capabilities of comparative genomics by tracing the evolution of additional biomolecular processes.

Structural similarity among proteins is reflected in the distribution of hydropathicity along the amino acids in the protein sequence. Similarities in the hydropathy distributions are obvious for homologous proteins within a protein family. They also were observed for proteins with related structures, even when sequence similarities were undetectable. Here we present a novel method that employs the hydropathy distribution in proteins for identification of (sub)families in a set of (homologous) proteins. We represent proteins as points in a generalized hydropathy space, represented by vectors of specifically defined features. The features are derived from hydropathy of the individual amino acids. Projection of this space onto principal axes reveals groups of proteins with related hydropathy distributions. The groups identified correspond well to families of structurally and functionally related proteins. We found that this method accurately identifies protein families in a set of proteins, or subfamilies in a set of homologous proteins. Our results show that protein families can be identified by the analysis of hydropathy distribution, without the need for sequence alignment., (Copyright 2005 Wiley-Liss, Inc.)

Methanotrophs are ubiquitous bacteria that can use the greenhouse gas methane as a sole carbon and energy source for growth, thus playing major roles in global carbon cycles, and in particular, substantially reducing emissions of biologically generated methane to the atmosphere. Despite their importance, and in contrast to organisms that play roles in other major parts of the carbon cycle such as photosynthesis, no genome-level studies have been published on the biology of methanotrophs. We report the first complete genome sequence to our knowledge from an obligate methanotroph, Methylococcus capsulatus (Bath), obtained by the shotgun sequencing approach. Analysis revealed a 3.3-Mb genome highly specialized for a methanotrophic lifestyle, including redundant pathways predicted to be involved in methanotrophy and duplicated genes for essential enzymes such as the methane monooxygenases. We used phylogenomic analysis, gene order information, and comparative analysis with the partially sequenced methylotroph Methylobacterium extorquens to detect genes of unknown function likely to be involved in methanotrophy and methylotrophy. Genome analysis suggests the ability of M. capsulatus to scavenge copper (including a previously unreported nonribosomal peptide synthetase) and to use copper in regulation of methanotrophy, but the exact regulatory mechanisms remain unclear. One of the most surprising outcomes of the project is evidence suggesting the existence of previously unsuspected metabolic flexibility in M. capsulatus, including an ability to grow on sugars, oxidize chemolithotrophic hydrogen and sulfur, and live under reduced oxygen tension, all of which have implications for methanotroph ecology. The availability of the complete genome of M. capsulatus (Bath) deepens our understanding of methanotroph biology and its relationship to global carbon cycles. We have gained evidence for greater metabolic flexibility than was previously known, and for genetic components that may have biotechnological potential., Competing Interests: The authors have declared that no conflicts of interest exist.

This work describes the development of a program that predicts whether or not a polypeptide sequence from a Gram-negative bacterium is an integral beta-barrel outer membrane protein. The program, called the beta-barrel Outer Membrane protein Predictor (BOMP), is based on two separate components to recognize integral beta-barrel proteins. The first component is a C-terminal pattern typical of many integral beta-barrel proteins. The second component calculates an integral beta-barrel score of the sequence based on the extent to which the sequence contains stretches of amino acids typical of transmembrane beta-strands. The precision of the predictions was found to be 80% with a recall of 88% when tested on the proteins with SwissProt annotated subcellular localization in Escherichia coli K 12 (788 sequences) and Salmonella typhimurium (366 sequences). When tested on the predicted proteome of E.coli, BOMP found 103 of a total of 4346 polypeptide sequences to be possible integral beta-barrel proteins. Of these, 36 were found by BLAST to lack similarity (E-value score < 1e-10) to proteins with annotated subcellular localization in SwissProt. BOMP predicted the content of integral beta-barrels per predicted proteome of 10 different bacteria to range from 1.8 to 3%. BOMP is available at http://www.bioinfo.no/tools/bomp.

Motivation: Many of the most interesting functional and evolutionary relationships among proteins are so ancient that they cannot be reliably detected through sequence analysis and are apparent only through a comparison of the tertiary structures. The conserved features can often be described as structural motifs consisting of a few single residues or Secondary Structure (SS) elements. Confidence in such motifs is greatly boosted when they are found in more than a pair of proteins., Results: We describe an algorithm for the automatic discovery of recurring patterns in protein structures. The patterns consist of individual residues having a defined order along the protein's backbone that come close together in the structure and whose spatial conformations are similar. The residues in a pattern need not be close in the protein's sequence. The work described in this paper builds on an earlier reported algorithm for motif discovery. This paper describes a significant improvement of the algorithm which makes it very efficient. The improved efficiency allows us to use it for doing unsupervised learning of patterns occurring in small subsets in a large set of structures, a non-redundant subset of the Protein Data Bank (PDB) database of all known protein structures.

Motivation: Radiation Hybrid Mapping (RHM) is a technique used to order a set of markers on a genome and estimating physical distances between them. RHM provides information on marker placement independent from other methods such as sequencing, and can therefore be used for example in genome sequencing to help ordering contigs. A radiation hybrid framework can be constructed by choosing a set of markers so that the chromosome coverage is good and so that the markers can be ordered with high confidence. Automatically constructing RHM frameworks is a computationally challenging problem., Results: We have developed a new method for constructing radiation hybrid frameworks. Given a relatively large set of markers for a chromosome, the algorithm aims to select an ordered subset that makes up a framework, and that contains as many markers as possible. The algorithm has a time complexity that is better than any of the existing methods that we are aware of. Furthermore, we propose a method for comparing if two frameworks are consistent, giving a visual presentation as well as quantitative measures of how well the two frameworks agree. Applying our method on marker sets from 22 human chromosomes and comparing the resulting frameworks with previously published frameworks, we demonstrate that our automatic method efficiently constructs frameworks with good coverage of each chromosome and with high degree of agreement on the marker ordering.

A method is described in which proteins that match PROSITE patterns are filtered by the root-mean-square deviation of the local 3D structures of the probe and target over the pattern components. This was found to increase the discrimination between true and false members of the protein family but was dependent on how unique the structural features in the pattern were compared to equivalent fragments extracted from the structure databank (for example; if the pattern fell in an alpha-helix, then discrimination was poor.) We then generalised the sequence patterns (by widening the range of amino acid residues allowed at each position) and monitored how well the structural information helped retain specificity. While the discrimination of the pure sequence pattern had generally disappeared at information content values less than ten bits, the discrimination of the combined sequence structure probe remained high at this point before following a similar decay. The displacement between these curves indicates that the structural component is, on average, equivalent to about ten bits. The sequence patterns were also filtered using the structure comparison program SAP, giving a global, rather than local "view" of the proteins. This allowed the information content of the sequence patterns to become even less specific but raised problems of whether some proteins encountered with the same fold but no PROSITE pattern should constitute family members., (Copyright 2000 Academic Press.)

A method of multiple sequence alignment is described based on the double dynamic programming (DDP) algorithm previously used for treating structural constraints encountered in structure comparison and threading. Following these applications, the inconsistencies that emerge when trying to combine pair-wise alignments into a multiple alignment are reconciled by summing all the, possibly inconsistent, paths (low-level alignments) into a matrix which is then used to provide a final (high-level) alignment. This process is applied to all sequence pairs and the pair-wise results combined in a simple multiple sequence alignment program. From this alignment, further constraints are selected to bias the low-level alignments in the DDP algorithm and the process iterated. The results, however, showed that this overall iteration was not needed and one-pass gave results at least as good as the 'standard' progressive method of multiple sequence alignment. Further applications of the method are discussed.

This article investigates aspects of pairwise and multiple structure comparison, and the problem of automatically discover common patterns in a set of structures. Descriptions and representation of structures and patterns are described, as well as scoring and algorithms for comparison and discovery. A framework and nomenclature is developed for classifying different methods, and many of these are reviewed and placed into this framework.

We present a language for describing structural patterns of residues in protein structures and a method for the discovery of such patterns that recur in a set of protein structures. The patterns impose restrictions on the spatial position of each residue, their order along the amino acid chain, and which amino acids are allowed in each position. Unlike other methods for comparing sets of protein structures, our method is not based on the use of pairwise structure comparisons which is often time consuming and can produce inconsistent results. Instead, the method simultaneously takes into account information from all structures in the search for conserved structure patterns which are potential structure motifs. The method is based on describing the spatial neighborhoods of each residue in each structure as a string and applying a sequence pattern discovery method to find patterns common to subsets of these strings. Finally it is checked whether the similarities between the neighborhood strings correspond to spatially similar substructures. We apply the method to analyze sets of very disparate proteins from the four different protein families: serine proteases, cuprodoxins, cysteine proteinases, and ferredoxins. The motifs found by the method correspond well to the site and motif information given in the annotation of these proteins in PDB, Swiss-Prot, and PROSITE. Furthermore, the motifs are confirmed by using the motif data to constrain the structural alignment of the proteins obtained with the program SAP. This gave the best superposition/alignment of the proteins given the motif assignment.

This paper surveys approaches to the discovery of patterns in biosequences and places these approaches within a formal framework that systematises the types of patterns and the discovery algorithms. Patterns with expressive power in the class of regular languages are considered, and a classification of pattern languages in this class is developed, covering the patterns that are the most frequently used in molecular bioinformatics. A formulation is given of the problem of the automatic discovery of such patterns from a set of sequences, and an analysis is presented of the ways in which an assessment can be made of the significance of the discovered patterns. It is shown that the problem is related to problems studied in the field of machine learning. The major part of this paper comprises a review of a number of existing methods developed to solve the problem and how these relate to each other, focusing on the algorithms underlying the approaches. A comparison is given of the algorithms, and examples are given of patterns that have been discovered using the different methods.