The Human EST Ontology Explorer: a tissue-oriented visualization system for ontologies distribution in human EST collectionsReportar como inadecuado

The Human EST Ontology Explorer: a tissue-oriented visualization system for ontologies distribution in human EST collections - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

BMC Bioinformatics

, 10:S2

First Online: 15 October 2009


BackgroundThe NCBI dbEST currently contains more than eight million human Expressed Sequenced Tags ESTs. This wide collection represents an important source of information for gene expression studies, provided it can be inspected according to biologically relevant criteria. EST data can be browsed using different dedicated web resources, which allow to investigate library specific gene expression levels and to make comparisons among libraries, highlighting significant differences in gene expression. Nonetheless, no tool is available to examine distributions of quantitative EST collections in Gene Ontology GO categories, nor to retrieve information concerning library-dependent EST involvement in metabolic pathways. In this work we present the Human EST Ontology Explorer HEOE est explorer, a web facility for comparison of expression levels among libraries from several healthy and diseased tissues.

ResultsThe HEOE provides library-dependent statistics on the distribution of sequences in the GO Direct Acyclic Graph DAG that can be browsed at each GO hierarchical level. The tool is based on large-scale BLAST annotation of EST sequences. Due to the huge number of input sequences, this BLAST analysis was performed with the aid of grid computing technology, which is particularly suitable to address data parallel task. Relying on the achieved annotation, library-specific distributions of ESTs in the GO Graph were inferred. A pathway-based search interface was also implemented, for a quick evaluation of the representation of libraries in metabolic pathways. EST processing steps were integrated in a semi-automatic procedure that relies on Perl scripts and stores results in a MySQL database. A PHP-based web interface offers the possibility to simultaneously visualize, retrieve and compare data from the different libraries. Statistically significant differences in GO categories among user selected libraries can also be computed.

ConclusionThe HEOE provides an alternative and complementary way to inspect EST expression levels with respect to approaches currently offered by other resources. Furthermore, BLAST computation on the whole human EST dataset was a suitable test of grid scalability in the context of large-scale bioinformatics analysis. The HEOE currently comprises sequence analysis from 70 non-normalized libraries, representing a comprehensive overview on healthy and unhealthy tissues. As the analysis procedure can be easily applied to other libraries, the number of represented tissues is intended to increase.

Electronic supplementary materialThe online version of this article doi:10.1186-1471-2105-10-S12-S2 contains supplementary material, which is available to authorized users.

Ivan Merelli, Andrea Caprera contributed equally to this work.

Download fulltext PDF

Autor: Ivan Merelli - Andrea Caprera - Alessandra Stella - Marcello Del Corvo - Luciano Milanesi - Barbara Lazzari


Documentos relacionados