doi:10.1093/bioinformatics/btk009BIOINFORMATICS ORIGINAL PAPER Databases and ontologies MiGenes: a searchable interspecies database of mitochondrial proteins curated using gene ontology annotation

Motivation: There has been an explosion of interest in the role of mitochondria in programmed cell death and other fundamental patho-logical processes underlying the development of human diseases. Nevertheless, the inventory of mitochondrial proteins encoded in the nuclear genome remains incomplete, providing an impediment to mito-chondrial researchat the interfacewith systemsbiology.Wecreated the MiGenes database to further define the scope of the mitochondrial proteome in humans and model organisms including mice, rats, flies and worms as well as budding and fission yeasts. MiGenes is intended to stimulate mitochondrial research using model organisms. Summary: MiGenes is a large-scale relational database that is auto-matically updated to keep pace with advances in mitochondrial proteomics and is curated to assure that the designation of proteins asmitochondrial reflects gene ontology (GO) annotations supported by high-quality evidence codes. A set of postulates is proposed to help define which proteins are authentic components of mitochondria. MiGenes incorporates>1160 new GO annotations to human, mouse and rat protein records, 370 of which represent the first GO annotation reflecting a mitochondrial localization. MiGenes employs a flexible search interface that permits batchwise accession number searches to support high-throughput proteomic studies. A web interface is pro-vided to permit members of the mitochondrial research community to suggest modifications in protein annotations or mitochondrial status. Availability: MiGenes is available a