Effect of utilizing terminology on extraction of protein-protein interaction information from biomedical literature

As the amount of on-line scientific literature in the biomedical domain increases, automatic processing has become a promising approach for accelerating research. We are applying syntactic parsing trained on the general domain to identify proteinprotein interactions. One of the main difficulties obstructing the use of language processing is the prevalence of specialized terminology. Accordingly, we have created a specialized dictionary by compiling on-line glossaries, and have applied it for information extraction. We conducted preliminary experiments on one hundred sentences, and compared the extraction performance when (a) using only a general dictionary and (b) using this plus our specialized dictionary. Contrary to our expectation, using only the general dictionary resulted in better performance (recall 93.0%, precision 91.0%) than with the terminology-based approach (recall 92.9%, precision 89.6%).