Extracting the structures of small molecules from unstructured text is now a mature field, however there still remain areas that present considerable difficulty or have until this point remained unexplored.One such area is identification of chemical names with misspellings or errors introduced by optical character recognition. The approach we have taken employs a formal grammar describing the syntax of a systematic name. To provide coverage over the vast majority of organic nomenclature including carbohydrates, amino acids and natural products we have developed a  new way of representing the grammar such as to allow an order of magnitude more states than previous efforts1 whilst  simultaneously reducing memory consumption. To efficiently perform spelling correction against this grammar we will describe a heuristic spelling correction algorithm.Another area that remains underexplored is the identification and resolution of chemical line formulae by which we also include domain specific line formulae such as are used to describe oligosaccharides and peptides. We describe the recognition and resolution of these often overlooked chemical entities.We also show how one can identify entities such as journal and patent references, which can aid in the navigation of semantically enhanced documents.(1) Sayle, R.; Xie, P. H.; Muresan, S. Improved Chemical Text Mining of Patents with Infinite Dictionaries and Automatic Spelling Correction. J. Chem. Inf. Model. 2011 , 52, 51�62.

Tackling the difficult areas of chemical entity extraction: Misspelt chemical names and unconventional entities

246th National Meeting (2013)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.