Geosemantically-enhanced PubMed queries using the GeoNames ontology and Web Services

PubMed is a popular biomedical bibliographic database curated by the US National Library of Medicine. Limitations may be encountered when querying PubMed for studies from or about a particular geographic location. For example, if a citation record does not include the word `Canada´ but instead includes `Ontario´ as the named place in its text, a PubMed query to find all studies conducted in Canada using only the keyword `Canada´ will not be able to retrieve that particular record. We propose that a GeoNames-powered PubMed search has the potential of solving such issues. By bringing in the “intelligence” of a geographic ontology and the ability to reason with place names (e.g., recognise `Ontario´ as a province of `Canada´), a search engine can infer implicit meanings in the queried records that are not directly mentioned as free text or as keyword terms in either the citation records or their metadata.