MOTIVATION: Assembling the relevant information needed to interpret the output from high-throughput, genome scale, experiments such as gene expression microarrays is challenging. Analysis reveals genes that show statistically significant changes in expression levels, but more information is needed to determine their biological relevance. The challenge is to bring these genes together with biological information distributed across hundreds of databases or buried in the scientific literature (millions of articles). Software tools are needed to automate this task which at present is labor-intensive and requires considerable informatics and biological expertise. RESULTS: This article describes ONDEX and how it can be applied to the task of interpreting gene expression results. ONDEX is a database system that combines the features of semantic database integration and text mining with methods for graph-based analysis. An overview of the ONDEX system is presented, concentrating on recently developed features for graph-based analysis and visualization. A case study is used to show how ONDEX can help to identify causal relationships between stress response genes and metabolic pathways from gene expression data. ONDEX also discovered functional annotations for most of the genes that emerged as significant in the microarray experiment, but were previously of unknown function.

BACKGROUND: Ontologies and taxonomies are among the most important computational resources for molecular biology and bioinformatics. A series of recent papers has shown that the Gene Ontology (GO), the most prominent taxonomic resource in these fields, is marked by flaws of certain characteristic types, which flow from a failure to address basic ontological principles. As yet, no methods have been proposed which would allow ontology curators to pinpoint flawed terms or definitions in ontologies in a systematic way. RESULTS: We present computational methods that automatically identify terms and definitions which are defined in a circular or unintelligible way. We further demonstrate the potential of these methods by applying them to isolate a subset of 6001 problematic GO terms. By automatically aligning GO with other ontologies and taxonomies we were able to propose alternative synonyms and definitions for some of these problematic terms. This allows us to demonstrate that these other resources do not contain definitions superior to those supplied by GO. CONCLUSION: Our methods provide reliable indications of the quality of terms and definitions in ontologies and taxonomies. Further, they are well suited to assist ontology curators in drawing their attention to those terms that are ill-defined. We have further shown the limitations of ontology mapping and alignment in assisting ontology curators in rectifying problems, thus pointing to the need for manual curation.

In the light of the increasing number of biological databases, their integration is a fundamental prerequisite for answering complex biological questions. Database integration, therefore, is an important area of research in bioinformatics. Since most of the publicly available life science databases are still exclusively exchanged by means of proprietary flat files, database integration requires parsers for very different flat file formats. Unfortunately, the development and maintenance of database specific flat file parsers is a nontrivial and time-consuming task, which takes considerable effort in large-scale integration scenarios. This paper introduces heuristically based concepts for automatic structure extraction from life science database flat files. On the basis of these concepts the FlatEx prototype is developed for the automatic conversion of flat files into XML representations.

Biology can be regarded as a science of networks: interactions between various biological entities (eg genes, proteins, metabolites) on different levels (eg gene regulation, cell signalling) can be represented as graphs and, thus, analysis of such networks might shed new light on the function of biological systems. Such biological networks can be obtained from different sources. The extraction of networks from text is an important technique that requires the integration of several different computational disciplines. This paper summarises the most important steps in network extraction and reviews common approaches and solutions for the extraction of biological networks from scientific literature.

Several hundred internet accessible life science databases with constantly growing contents and varying areas of specialization are publicly available via the internet. Database integration, consequently, is a fundamental prerequisite to be able to answer complex biological questions. Due to the presence of syntactic, schematic, and semantic heterogeneities, large scale database integration at present takes considerable efforts. As there is a growing apprehension of extensible markup language (XML) as a means for data exchange in the life sciences, this article focuses on the impact of XML technology on database integration in this area. In detail, a general architecture for ontology-driven data integration based on XML technology is introduced, which overcomes some of the traditional problems in this area. As a proof of concept, a prototypical implementation of this architecture based on a native XML database and an expert system shell is described for the realization of a real world integration scenario.

MOTIVATION: Many molecular biological databases are implemented on relational Database Management Systems, which provide standard interfaces like JDBC and ODBC for data and metadata exchange. By using these interfaces, many technical problems of database integration vanish and issues related to semantics remain, e.g. the use of different terms for the same things, different names for equivalent database attributes and missing links between relevant entries in different databases. RESULTS: In this publication, principles and methods that were used to implement SEMEDA (Semantic Meta Database) are described. Database owners can use SEMEDA to provide semantically integrated access to their databases as well as to collaboratively edit and maintain ontologies and controlled vocabularies. Biologists can use SEMEDA to query the integrated databases in real time without having to know the structure or any technical details of the underlying databases. AVAILABILITY: SEMEDA is available at http://www-bm.ipk-gatersleben.de/semeda/. Database providers who intend to grant access to their databases via SEMEDA are encouraged to contact the authors.

Two fundamental approaches for database integration exist: the data warehouse approach attempts to physically merge data sets from several source databases, whereas database federations simultaneously query source databases online. In this paper, a database federation approach based on two components is introduced. MARGBench is a system which, among other features, enables the querying of several databases in SQL by translating SQL queries into a source database-specific interface. In this system, SQL queries use the database field labels of the original database, i.e. fields that contain the same kind of information are labelled differently in different databases. In order to solve this problem, a second system, based on ontologies, is currently being developed. This ontology system not only includes information about the semantics of database fields but also contains information about the databases themselves. Thus, it facilitates both database binding and intelligent database querying. The main concepts and ideas of these two systems are explained. By using an imaginary database query, it is demonstrated how the two systems (the ontology system and MARGBench) work together in order to enable the querying of several databases simultaneously.