Skip to main content
. Author manuscript; available in PMC: 2018 Oct 1.
Published in final edited form as: Int J Med Inform. 2017 Jul 24;106:48–56. doi: 10.1016/j.ijmedinf.2017.07.002

Table 6.

Datasets, data analysis methods, and key findings for selected studies in research foci 4 - mining the literature for knowledge discovery.

Study Datasets Primary data analysis methods Key findings
Malhotra et al. (54) 3,007 MEDLINE abstracts, and annotations of non-overlapping 200 randomly chosen abstracts HypothesisFinder (including dictionary based named entity recognition and rule-based pattern recognition) HypothesisFinder can build hypothetical protein interaction network.
Greco et al. (55) Full text literature reviews, MEDLINE, web-based reports, and databases (e.g., gene expression databases, protein-pathway databases, protein-disease association databases) Text mining 25 potential candidate biomarkers were found. Two candidate biomarkers (i.e., Choline Acetyltransferase (ChAt) and urokinase-type Plasminogen Activator Receptor (PLAUR)) have not been reported previously.
Li et al. (56) Online Predicted Human Interaction Database (OPHID) for protein interaction, 222,609 AD-related MEDLINE abstracts for AD drugs Molecular interaction network mining, text mining 17 candidate AD drugs were found
Jonnalagadda et al. (57) MEDLINE abstracts Information extraction and semantic information extraction 336 AD-related sentences from 194 abstracts in the MEDLINE, 84.5% of which were evaluated as relevant abstracts