GOLink: Finding Cooccurring Terms across Gene Ontology Namespaces.

The Gene Ontology (GO) provides a resource for consistent annotation of genes and gene products that is extensively used by numerous large public repositories. The GO is constructed of three subontologies describing the cellular component of action, molecular function, and overall biological process of a gene or gene product. Querying across the subontologies is problematic and no standard method exists to, for example, find all molecular functions occurring in a particular cellular component. GOLink addresses this problem by finding terms from all subontologies cooccurring with a term of interest in annotation across the entire GO database. Genes annotated with this term are exported and their GO annotation is assigned to three separate GOLink terms lists based on specific criteria. The software was used to predict the most likely Biological Process for a group of genes using just their Molecular Function terms giving sensitivity, specificity, and accuracy between 80 and 90% across all the terms lists. GOLink is made freely available for noncommercial use and can be downloaded from the project website.