PubMed® Corrects Spelling

Entrez’s PubMed database now includes a spell-check feature that offers spelling corrections for words within PubMed queries that appear to be misspelled. For instance, a misspelled query such as “celular” generates a little over 4,600 hits. However, PubMed now returns a link to the results of a correctly spelled query:

Did you mean: cellular (341,678 items)

Access to the PubMed spell-checker via scripts is provided by a new Entrez Utility called “espell”. Espell takes four parameters; “db”, “term”, “email” and “tool”. For example, an Eutility call such as:

eutils.ncbi.nlm.nih.gov/entrez/eutils/espell.fcgi?db=pubmed&term=cardiiac+
thromboosis+ischemia

returns the following XML formatted result:

<?xml version=”1.0”?> <!DOCTYPE eSpellResult PUBLIC “-//NLM//DTD eSpellResult, 23 November 2004//EN” “http://www.ncbi.nlm.nih.gov/entrez/query/DTD/eSpell.dtd”> <eSpellResult>   <Database>pubmed</Database> <Query>cardiiac thromboosis ischemia</Query> <CorrectedQuery>cardiac thrombosis ischemia</CorrectedQuery> <SpelledQuery><Replaced>cardiac</Replaced><Original> </Original><Replaced>thrombosis</Replaced><Original> ischemia</Original></SpelledQuery>   <ERROR/> </eSpellResult>

In the call above, the “db” parameter specifies the “PubMed” database, the only database supported at present, while the “term” parameter gives the search phrase. As with all Eutility calls, spaces within search phrases are represented with “+” signs. The parameters “email” and “tool” are optional, however you may use them to provide an email address that NCBI can use to contact you and to identify your script, respectively. The use of the “email” and “tool” parameters is helpful in cases of a script malfunction.

The result includes the original query, the complete corrected query, and a breakdown of the terms in the complete corrected query flagged as either “replaced” or “original”.

Terms that are field-restricted, such as “heart[title]”, are not checked since each field implies the use of a separate vocabulary that may include standard abbreviations that would be flagged as misspelled words in the context of a more general usage.

To read more about the Entrez Utilities, or to subscribe to the “utilities-announce” mailing list, see:

eutils.ncbi.nlm.nih.gov/entrez/query/static/eutils_help.html