An ontology for bioinformatics applications

MOTIVATION

An ontology of biological terminology provides a model of biological concepts that can be used to form a semantic framework for many data storage, retrieval and analysis tasks. Such a semantic framework could be used to underpin a range of important bioinformatics tasks, such as the querying of heterogeneous bioinformatics sources or the systematic annotation of experimental results.

RESULTS

This paper provides an overview of an ontology [the Transparent Access to Multiple Biological Information Sources (TAMBIS) ontology or TaO] that describes a wide range of bioinformatics concepts. The present paper describes the mechanisms used for delivering the ontology and discusses the ontology's design and organization, which are crucial for maintaining the coherence of a large collection of concepts and their relationships.

AVAILABILITY

The TAMBIS system, which uses a subset of the TaO described here, is accessible over the Web via http://img.cs.man.ac.uk/tambis (although in the first instance, we will use a password mechanism to limit the load on our server). The complete model is also available on the Web at the above URL.

Metaphyseal chondrodysplasia type Schmid mutations are predicted to occur in two distinct three-dimensional clusters within type X collagen NC1 domains that retain the ability to trimerize

Metaphyseal chondrodysplasia type Schmid (MCDS) is caused by mutations in COL10A1 that are clustered in the carboxyl-terminal non-collagenous (NC1) encoding domain. This domain is responsible for initiating trimerization of type X collagen during biosynthesis. We have built a molecular model of the NC1 domain trimer based on the crystal structure coordinates of the highly homologous trimeric domain of ACRP30 (adipocyte complement-related protein of 30 kDa or AdipoQ). Mapping of the MCDS mutations onto the structure reveals two specific clusters of residues as follows: one on the surface of the monomer which forms a tunnel through the center of the assembled trimer and the other on a patch exposed to solvent on the exterior surface of each monomeric unit within the assembled trimer. Biochemical studies on recombinant trimeric NC1 domain show that the trimer has an unusually high stability not exhibited by the closely related ACRP30. The high thermal stability of the trimeric NC1 domain, in comparison with ACRP30, appears to be the result of a number of factors including the 17% greater total buried solvent-accessible surface and the increased numbers of hydrophobic contacts formed upon trimerization. The 27 amino acid sequence present at the amino terminus of the NC1 domain, which has no counterpart in ACRP30, also contributes to the stability of the trimer. We have also shown that NC1 domains containing the MCDS mutations Y598D and S600P retain the ability to homotrimerize and heterotrimerize with wild type NC1 domain, although the trimeric complexes formed are less stable than those of the wild type molecule. These studies suggest strongly that the predominant mechanism causing MCDS involves a dominant interference of mutant chains on wild type chain assembly.

A RAPID algorithm for sequence database comparisons: application to the identification of vector contamination in the EMBL databases

MOTIVATION

Word-matching algorithms such as BLAST are routinely used for sequence comparison. These algorithms typically use areas of matching words to seed alignments which are then used to assess the degree of sequence similarity. In this paper, we show that by formally separating the word-matching and sequence-alignment process, and using information about word frequencies to generate alignments and similarity scores, we can create a new sequence-comparison algorithm which is both fast and sensitive. The formal split between word searching and alignment allows users to select an appropriate alignment method without affecting the underlying similarity search. The algorithm has been used to develop software for identifying entries in DNA sequence databases which are contaminated with vector sequence.

RESULTS

We present three algorithms, RAPID, PHAT and SPLAT, which together allow vector contaminations to be found and assessed extremely rapidly. RAPID is a word search algorithm which uses probabilities to modify the significance attached to different words; PHAT and SPLAT are alignment algorithms. An initial implementation has been shown to be approximately an order of magnitude faster than BLAST. The formal split between word searching and alignment not only offers considerable gains in performance, but also allows alignment generation to be viewed as a user interface problem, allowing the most useful output method to be selected without affecting the underlying similarity search. Receiver Operator Characteristic (ROC) analysis of an artificial test set allows the optimal score threshold for identifying vector contamination to be determined. ROC curves were also used to determine the optimum word size (nine) for finding vector contamination. An analysis of the entire expressed sequence tag (EST) subset of EMBL found a contamination rate of 0.27%. A more detailed analysis of the 50 000 ESTs in est10.dat (an EST subset of EMBL) finds an error rate of 0.86%, principally due to two large-scale projects.

AVAILABILITY

A Web page for the software exists at http://bioinf.man.ac.uk/rapid, or it can be downloaded from ftp://ftp.bioinf.man.ac.uk/RAPID CONTACT: crispin@cs.man.ac.uk

INTERACT: an object oriented protein-protein interaction database

MOTIVATION

Protein-protein interactions provide vital information concerning the function of proteins, complexes and networks. Currently there is no widely accepted repository of this interaction information. Our aim is to provide a single database with the necessary architecture to fully store, query and analyse interaction data.

RESULTS

An object oriented database has been created which provides scientists with a resource for examining existing protein-protein interactions and inferring possible interactions from the data stored. It also provides a basis for examining networks of interacting proteins, via analysis of the data stored. The database contains over a thousand interactions.

CONTACT

k.eilbeck@stud.man.ac.uk

Deducing polymeric structure from aqueous molecular dynamics simulations of oligosaccharides: predictions from simulations of hyaluronan tetrasaccharides compared with hydrodynamic and X-ray fibre diffraction data

Molecular dynamics simulations of the two hyaluronan tetrasaccharides in water predict that over a period of 500 ps, their central linkages populate a single primary minima. Over the same period the peripheral linkages explore this minima, but also a secondary minima. Structures constructed using the primary minima were found to be extended left-handed helices of axial rise per disaccharide (h) 0.8 to 1.0 nm and 2.8 to 4.5 disaccharides per turn (n), in good agreement with n=3 and n=4 helices found by X-ray fibre diffraction studies. We have used the predicted average conformation from molecular dynamics to calculate the translational diffusion coefficients of the oligosaccharide series up to decasaccharide, and compared these with experimental measurements obtained using the method of capillary dispersion. Our calculated values are found to be in good agreement with experiment beyond the size of a tetrasaccharide. A partial digest of hyaluronan in the molecular mass range 10 to 100 kDa was fractionated by gel chromatography. Molecular weights were determined by in-line laser light-scattering measurements, and the translational diffusion coefficients of selected fractions were determined by dynamic laser light-scattering. A similar experiment was performed on hyaluronan with a molecular mass greater than 1MDa. The data suggest a change from rod-like to stiff coil behaviour beyond a molecular weight of 10 kDa. We have also examined the conformations available using the secondary minima, found at the peripheral linkages. In contrast to the extended structures previously described we have found left and right-handed helices with high values of n (5-10) and low values of h. Although there is no experimental evidence for these structures, they are of interest as, over short stretches, they would introduce folds, loops, and turns into the hyaluronan molecule. Such shapes may play an important role in the hydrodynamics of hyaluronan and its interaction with lipids and proteins.

TAMBIS--Transparent Access to Multiple Bioinformatics Information Sources

The TAMBIS project aims to provide transparent access to disparate biological databases and analysis tools, enabling users to utilize a wide range of resources with the minimum of effort. A prototype system has been developed that includes a knowledge base of biological terminology (the biological Concept Model), a model of the underlying data sources (the Source Model) and a 'knowledge-driven' user interface. Biological concepts are captured in the knowledge base using a description logic called GRAIL. The Concept Model provides the user with the concepts necessary to construct a wide range of multiple-source queries, and the user interface provides a flexible means of constructing and manipulating those queries. The Source Model provides a description of the underlying sources and mappings between terms used in the sources and terms in the biological Concept Model. The Concept Model and Source Model provide a level of indirection that shields the user from source details, providing a high level of source transparency. Source independent, declarative queries formed from terms in the Concept Model are transformed into a set of source dependent, executable procedures. Query formulation, translation and execution is demonstrated using a working example.

Dynamic exchange between stabilized conformations predicted for hyaluronan tetrasaccharides: comparison of molecular dynamics simulations with available NMR data

Studies of the hyaluronan (HA) tetrasaccharides are important for understanding hydrogen-bonding in the HA polymer, as they are probably the smallest oligomers in which characteristics of the constituent monosaccharides and the polymer are simultaneously exhibited. Here we present extensive molecular dynamics simulations of the two tetrasaccharides of HA in dilute aqueous solution. These simulations have confirmed the existence of intramolecular hydrogen-bonds between the neighboring sugar residues of HA in solution, as proposed by Scott (1989). However, our simulations predict that these intramolecular hydrogen-bonds are not static as previously proposed, but are in constant dynamic exchange on the sub-nanosecond time-scale. This process results in discrete internal motion of the HA tetrasaccharides where they rapidly move between low energy conformations. Specific interactions between water and intramolecular hydrogen-bonds involving the hydroxymethyl group were found to result in differing conformations and dynamics for the two alternative tetrasaccharides of HA. This new observation suggests that this residue may play a key role in the entropy and stability of HA in solution, allowing it to stay soluble up to high concentration. The vicinal coupling constants3 J NHCH of the acetamido groups have been calculated from our aqueous simulations of HA. We found that high values of 3J NHCH approximately 8 Hz, as experimentally measured for HA, are consistent with mixtures of both trans and cis conformations, and thus3 J NHCH cannot be used to imply a purely trans conformation of the acetamido. The rapid exchange of intramolecular hydrogen-bonds indicates that although the structure is at any moment stabilized by these hydrogen-bonds, no one hydrogen-bond exists for an extended period of time. This could explain why NMR often fails to provide evidence for intramolecular hydrogen-bonds in HA and other aqueous carbohydrate structures.

Searching DNA databases for similarities to DNA sequences: when is a match significant?

MOTIVATION

Searching DNA sequences against a DNA database is an essential element of sequence analysis. However, few systematic studies have been carried out to determine when a match between two DNA sequences has biological significance and this is limiting the use that can be made of DNA searching algorithms.

RESULTS

A test set of DNA sequences has been constructed consisting of artificially evolved and real sequences. This set has been used to test various database searching algorithms (BLAST, BLAST2, FASTA and Smith-Waterman) on a subset of the EMBL database. The results of this analysis have been used to determine the sensitivity and coverage of all of the algorithms. Guidelines have been produced which can be used to assess the significance of DNA database search results. The Smith-Waterman algorithm was shown to have the best coverage, but the worst sensitivity, whereas the default BLASTN algorithm (word length set to 11) was shown to have good sensitivity, but poor coverage. A sensible compromise between speed, sensitivity and coverage can be obtained using either the FASTA or BLAST (word length set to 6) algorithms. However, analysis of the results also showed that no algorithm works well when the length of the probe sequence is <200 bases. In general, matches can accurately be identified between coding regions of DNA sequences when there is >35% sequence identity between the corresponding proteins. Searching a DNA sequence against a DNA sequence database can, therefore, be a useful tool in sequence analysis.

AVAILABILITY

The test sets used are available via anonymous ftp from mbisg2.sbc.man.ac.uk in the directory /pub/cabios/testdata/

CONTACT

I.Anderson@stud.man.ac.uk; abrass@man.ac.uk

Recent developments in biological sequence databases

Biological sequence databases are currently being re-engineered to make them more efficient and easier to use. This re-engineering is also providing an infrastructure to make it easier to interrogate and integrate data from different sources. The net result of this effort should be a great improvement in the power and availability of bioinformatics resources to the general biology community.

The Ov20 protein of the parasitic nematode Onchocerca volvulus. A structurally novel class of small helix-rich retinol-binding proteins

Ov20 is a major antigen of the parasitic nematode Onchocerca volvulus, the causative agent of river blindness in humans, and the protein is secreted into the tissue occupied by the parasite. DNA encoding Ov20 was isolated, and the protein was expressed in Escherichia coli. Fluorescence-based ligand binding assays show that the protein contains a high affinity binding site for retinol, fluorescent fatty acids (11-((5-dimethylaminonaphthalene-1-sulfonyl)amino)undecanoic acid, dansyl-DL-alpha-aminocaprylic acid, and parinaric acid) and, by competition, oleic and arachidonic acids, but not cholesterol. The fluorescence emission of dansylated fatty acids is significantly blue-shifted upon binding in comparison to similarly sized beta-sheet-rich mammalian retinol- and fatty acid-binding proteins. Secondary structure prediction algorithms indicate that a alpha-helix predominates in Ov20, possibly in a coiled coil motif, with no evidence of beta structures, and this was confirmed by circular dichroism. The protein is highly stable in solution, requiring temperatures in excess of 90 degrees C or high denaturant concentrations for unfolding. Ov20 therefore represents a novel class of small retinol-binding protein, which appears to be confined to nematodes. The retinol binding activity of Ov20 could possibly contribute to the eye defects associated with onchocerciasis and, because there is no counterpart in mammals, represents a strategic target for chemotherapy.

Molecular dynamics simulations of the two disaccharides of hyaluronan in aqueous solution

Hyaluronan is an unusually stiff polymer when in aqueous solution, which has important consequences for its biological function. Molecular dynamics simulations of hyaluronan disaccharides have been performed, with explicit inclusion of water, to determine the molecular basis of this stiffness, and to investigate the dynamics of the glycosidic linkages. Our simulations reveal that stable sets of hydrogen bonds frequently connect the neighboring residues of hyaluronan. Water caging around the glycosidic linkage was observed to increase the connectivity between sugars, and further constrain them. This, we propose, explains the unusual stiffness of polymeric hyaluronan. It would allow the polysaccharide to maintain local secondary structure, and occupy large solution domains consistent with the visco-elastic nature of hyaluronan. Simulations in water showed no significant changes on inclusion of the exoanomeric effect. This, we deduced, was due to hyaluronan disaccharides ordering first shell water molecules. In some cases these waters were observed to transiently induce conformational change, by breaking intramolecular hydrogen bonds.

The mechanical properties of simulated collagen fibrils

Previous theoretical studies of the mechanical properties of tissues such as skin, bone and tendon, have used approaches based on composite materials and have tended to neglect the contribution of individual microscopic components. In this paper, we examine the relationship between the fine structure of a collagen fibril and its relative tensile strength. Collagen is a fibrous protein which provides associated tissues with the majority of their tensile strength. It is present in the form of elongated structures termed fibrils which are created by the self-assembly of rod-like collagen molecules in an entropy-driven process termed fibrillogenesis. Mutations that alter the primary structure of the collagen molecule, interfere with this assembly process and can lead to the potentially fatal brittle bone disease, osteogenesis imperfecta. Here we investigate the mechanical properties of a range of computer-generated aggregates. The aggregates, created by the diffusion limited aggregation of rods, were subjected to a simple tensile test based on local rules of damage accumulation. In the test, core samples are "extracted' from the aggregates, and the network of particles involved in the transmission of stress resolved. Increasing stress applied to the core leads to the removal of individual rods from this network; the tensile strength is determined from the force necessary to form a discontinuous network. Using this approach, we have shown that collagen fibril morphology is critical in determining its tensile strength. We suggest a possible mechanism to account for the increasing severity of osteogenesis imperfecta associated with the distance of mutation from the N-terminal of the collagen molecule.

Identification of the complete coding sequence and genomic organization of the Treacher Collins syndrome gene

Treacher Collins syndrome (TCS) is an autosomal dominant disorder of craniofacial development, the features of which include conductive hearing loss and cleft palate. Recently, the demonstration of a series of 10 mutations within a partial-length cDNA clone have indicated that the TCS gene (TCOF1) has been positionally cloned. Although it has been shown that the gene is expressed in a wide variety of fetal and adult tissues, database sequence comparisons have failed to provide significant information on the function of the gene. In the current investigation, a combination of cDNA library screening and rapid amplification of cDNA ends has permitted the isolation of the complete coding sequence of TCOF1, which is encoded by 26 exons and predicts a low complexity, serine/alanine-rich protein of approximately 144 kD. The use of a variety of bioinformatics tools has resulted in the identification of repeated units within the gene, each of which maps onto an individual exon. The predicted protein Treacle contains numerous potential phosphorylaiton sites, a number of which map to similar positions within the repeated units, and shows weak but significant homology to the nucleolar phosphoproteins. Although the precise function of Treacle remains unknown, these observations suggest that phosphorylation may be important for its role in early embryonic development and that it may play a role in nucleolar-cytoplasmic shuttling. The information presented in this study will allow continued mutation analysis in families with a history of TCS and should facilitate continued experimentation to shed further light on the function of the gene/protein during development of the craniofacial complex.

Protein secondary structure prediction by the analysis of variation and conservation in multiple alignments

A number of methods exist for the prediction of protein secondary structure from primary sequence. One method identifies variable charged and conserved hydrophobic residues within large multiple alignments as a means of indicating outside and inside sites respectively in the protein structure. These sites are then manually fitted to secondary structure templates to generate a secondary structure prediction. Using the existing theoretical bases of this method, we present an algorithm (STAMA) which automatically carries out the initial variation/conservation analysis of the alignment. We also test the accuracy of complete predictions carried out by manual fitting of the STAMA-derived assignments to structure templates, using five large multiple alignments each including a protein of known structure. The method was found on average to predict only 57% of residues in the correct secondary structure, and was only as accurate as predictions carried out using the established and automated method of Garnier, Osguthorpe and Robson (1978) applied to a single sequence. When used in conjunction with other secondary structure prediction methods, however, the resulting consensus predictions were found to be very accurate, with 78% of the elements (alpha helices or beta strands) for which a consensus could be obtained being predicted correctly. The algorithm presented here, plus the assessment of the accuracy of prediction generated by this method, should enable this predictive approach to receive informed general use.

The ABA-1 allergen of the parasitic nematode Ascaris suum: fatty acid and retinoid binding function and structural characterization

We report here on the structure and function of the ABA-1 allergen protein of the parasitic nematode Ascaris, the first nematode allergen to be characterized in detail. Using the fluorescent fatty acid analog 11-(((5-(dimethylamino)-1-naphthalenyl)sulfonyl)amino)undecanoic acid (DAUDA), it was demonstrated that ABA-1 is a fatty acid binding protein (FABP) with a high affinity for the fluorescent analog (8.8 x 10(-8) M) and for oleic acid in competition experiments (1.3 x 10(-8) M), with a single binding site for ligand per monomer unit. Blue-shifting of fluorescence emission of DAUDA upon binding was unprecedented in degree among FABPs, being equivalent to that occurring in cyclohexane. A similarly blue-shifted spectrum was obtained with a probe in which the fluorophore was bound to the alpha carbon of a fatty acid, indicating that the carboxylate group of bound fatty acids is probably not exposed to solvent. In competition experiments and by observation of changes in their intrinsic fluorescence, retinol and retinoic acid were also found to bind in the fatty acid binding site. Circular dichroism (CD) of the ABA-1 protein revealed a high alpha-helix content (59%) which was consistent with the four-helix structure for the protein predicted from sequence algorithms. Fluorescence measurements showed that the single, highly conserved tryptophan residue is deeply buried in an unusually apolar environment and that this was unaffected by ligand binding. DSC studies of thermal stability indicate that unfolding of the ABA-1 dimer is cooperative and biphasic (Tm approximately 71 and 89 degrees C), suggesting a two-domain thermal unfolding process, again consistent with the predicted structure. Only the folding of the high-temperature domain is reversible on cooling. DSC confirmed the gel filtration analysis, which indicated that ABA-1 forms a dimer. Aside from being the first nematode allergen for which structure or function has been elucidated, ABA-1 provides a highly manipulable model for investigation of the interaction between hydrophobic ligands and alpha-helical proteins.

Simple physical model of collagen fibrillogenesis based on diffusion limited aggregation

Type I collagen is a rod-like protein which self-assembles in a regular array to form elongated fibrils. The process of fibril formation, termed fibrillogenesis, is driven by the increase in entropy associated with loss of water from the bound monomers. A model based on diffusion limited aggregation (DLA) was used to investigate some of the mechanisms involved in this process. The aggregates created in the model displayed several features in common with collagen fibrils including an elongated morphology and a preference for tip growth. Analysis of these aggregates revealed a linear relationship between mass and distance from the tip, consistent with experimental observations. Intrafibrillar fluidity was introduced into the model by using a surface diffusion term. This led to the formation of aggregates with more compact morphologies. These results strongly implicate the role of diffusion limited growth in collagen fibril formation.

A secondary structure model of the integrin alpha subunit N-terminal domain based on analysis of multiple alignments

The integrins are alpha/beta heterodimeric proteins which mediate cell-matrix and cell-cell interactions. Current data indicate that the N-terminal moiety of the alpha subunit is involved in ligand binding. This region of the receptor is made up of a seven-fold repeated sequence of unknown structure which contains EF-hand-like putative divalent cation-binding sites. Recent studies have shown that multiple sequence alignments can be analysed to yield secondary structure predictions. Therefore, to obtain a model structure for the integrin alpha subunit N-terminal domain repeat, a large alignment of the seven repeats from sixteen integrin sequences was generated. Two methods of analysis were used: First, Chou and Fasman and Garnier, Osguthorpe and Robson predictions were carried out for individual sequences and the consensus predictions derived. Consensus hydrophobicity and chain flexibility data were also used to provide additional data. Second, sites of conservation and variation were analysed by a computer program STAMA (STructure After Multiple Alignment) to yield a secondary structure prediction. The two analyses gave essentially the same predicted structure: undefined region, loop, alpha-helix, beta-strand, divalent cation-binding loop, beta-strand, putative turn, loop, beta-strand. This is the first model structure to be presented for an integrin domain. Its implications for integrin function are discussed.

Identification of a key integrin-binding sequence in VCAM-1 homologous to the LDV active site in fibronectin

The integrin adhesion receptor alpha 4 beta 1 binds two ligands, the extracellular matrix glycoprotein fibronectin and the immunoglobulin superfamily member VCAM-1. Ligand-binding sites are contained with the HepII/IIICS domain of fibronectin, and within the homologous immunoglobulin domains 1 and 4 of VCAM-1. Previous studies have shown that the binding of each ligand to alpha 4 beta 1 is mutually exclusive, suggesting that they may employ similar mechanisms to bind receptor. Fibronectin contains at least three distinct peptide sequences that are active sites for alpha 4 beta 1 binding, two homologous sequences Leu-Asp-Val-Pro (LDVP) and Ile-Asp-Ala-Pro (IDAP), and a third related to Arg-Gly-Asp (RGD). Using a combination of site-directed mutagenesis and synthetic peptide approaches in conjunction with VCAM-1-dependent cell adhesion assays, we now report the identification of a key alpha 4 beta 1-binding sequence in both domains 1 and 4 of VCAM-1 as the tetrapeptide Ile-Asp-Ser-Pro (IDSP). Mutagenesis studies also suggest that an additional sequence in domain 1, KLEK, participates in receptor binding. Since IDSP is homologous to the LDVP and IDAP fibronectin peptides, this therefore provides a molecular explanation for the promiscuity of ligand binding by alpha 4 beta 1 and has implications for the design of synthetic VCAM-1 antagonists. The extrapolation of these findings to other integrin-binding immunoglobulin ligands is also discussed.