Search

In biomedical research, interpretation of microarray data requires confrontation of data and knowledge from heterogeneous resources, either in the biomedical domain or in genomics, as well as restitution and analysis methods adapted to huge amounts of data. We present a combined approach that relies on two components: BioMeKE annotates sets of genes using biomedical GO and UMLS concepts, and GEDAW, a Gene Expression Data Warehouse, uses BioMeKE to enrich experimental results with biomedical concepts, thus performing complex analyses of expression measurements through analysis workflows. The strength of our approach has been demonstrated within the framework of analysis of data resulting from the liver transcriptome study. It allowed new genes potentially associated with liver diseases to be highlighted.

Class 1 integrons are widespread genetic elements responsible for dissemination of antibiotic resistance among Gram-negative bacteria. Integrons allow bacteria to capture and express gene cassettes (GCs) via an integrase (IntI1) and a promoter (Pc) contained in the integron functional platform. GCs are transcribed from Pc, of which 13 variants of different strengths have been described, or, occasionally, from both Pc and a second promoter (P2). The intI1 promoter (PintI1) is repressed by LexA, which is the transcriptional repressor of the global regulatory SOS response network. Moreover, PintI1 lies face to face with Pc and overlaps P2, both configurations being propitious to transcriptional interference (TI). In this study, we analyzed possible transcriptional interference by quantifying transcripts produced from Pc, P2, and PintI1. We found that the Pc promoter interferes with the level of intI1 transcription but that this effect depends on the Pc variant: the strong Pc variant prevents intI1 expression, in contrast to the other variants. Although P2 formation results in LexA binding site disruption and thus prevents SOS regulation of intI1 expression, P2 does not interfere with PintI1. These findings reveal a tight relationship between GC and integrase expression. [ABSTRACT FROM AUTHOR]

Semantic interoperability between knowledge bases in medicine, and knowledge base in genomics and molecular biology will lead to advances in fundamental research as well as to improved patient care. DNA chips strategy is used for transcriptome analysis in order to identify deregulated genes in physio-pathological conditions. The objective of the BioMedical Knowledge Extraction project (BioMeKe) is to develop a knowledge warehouse in the context of transcriptome analysis during liver diseases. Knowledge sources include ontologies, related terminologies and annotations linked towards public databases (e.g., SWISSPROT). BioMeKe has been developed to have access to information using systematic investigation upon a concept, gene, gene products, pathology, or any target keyword, and is based on the combination of several relevant resources: UMLS, GeneOntology, MeSH supplementary terms, GOA, and HUGO. Current efforts are focusing on exploiting this ontology-based Knowledge Extractor, to enrich the expression data on genes delivered by a liver specific DNA microarray for better assistance of analysis.