The Alignment of the Medical Subject Headings to the Gene Ontology and Its Application in Gene Annotation

摘要

The Gene Ontology (GO) is a controlled vocabulary used for annotation of genes. Assigning such terms to uncategorized genes is time-consuming work, and a recurring task in biomedicine. The biomedical citations of the literature database MEDLINE are indexed with terms from the Medical Subject Headings (MeSH). We studied whether MeSH terms from gene-related MEDLINE entries could be translated to GO, and be used automatically for annotation purposes. We explored three MeSH to GO alignments: pairing similar MeSH and GO term synonyms, indirect linking through MeSH's associations with Enzyme Commission (EC) terms and the official EC2GO translation table, and using association analysis to indicate which MeSH and GO terms that co-occurred most frequently for existing annotations. We here show that an alignment can be found, and, despite inconsistency in the use of MeSH terms as MEDLINE indexes, we conclude that GO annotation prediction using this alignment is useful in manual annotation of genes.