A Graph-Based Recovery and Decomposition of Swanson's Hypothesis using Semantic Predications

TitleA Graph-Based Recovery and Decomposition of Swanson's Hypothesis using Semantic Predications
Publication TypeMiscellaneous
Year of Publication2012
AuthorsDelroy Cameron, Olivier Bodenreider, Hima Yalamanchili, Tu Danh, Sreeram Vallabhaneni, Krishnaprasad Thirunarayan, Amit Sheth, Thomas C. Rindflesch
Keywordss Hypothesis and Literature-Based Discovery (LBD), Semantic Associations and Background Knowledge and Semantic Predications and Subgraph Creation and Swanson&#39

Objectives: This paper proposes a novel graph-based framework that leverages semantic representations to recover and decompose Swanson's Raynaud Syndrome-Fish Oil Hypothesis. Our framework suggests that the future of Literature-based Discovery (LBD) depends heavily on the ability to semantically integrate information from scientific literature (called semantic predications) with information represented in structured background knowledge sources. Methods: The information extracted from the scientific literature forms the basis for the construction of labeled directed graphs, from which important associations among concepts have been identified. By manually clustering these associations to form expressive subgraphs, and then intuitively including background knowledge to bridge knowledge gaps in the literature, associations originally articulated by Swanson have been recaptured. Results: Using this graph-based framework, not only have we recovered the 3 associations commonly recognized as Swanson's Hypothesis, but we have also recovered an additional 11 out of 16 detailed associations that further explicate these known relations. To the best of our knowledge, such an in-depth decomposition of Swanson's Hypothesis has never been attempted. Conclusion: We therefore showed that a graph-based framework that leverages semantic representations is effective for recovering and decomposing Swanson's Hypothesis. In addition, we established that the future of LBD warrants: 1) expressive representations beyond Swanson's ABC model; 2) an ability to accurately extract semantic information from text; and 3) the semantic integration of scientific literature and background knowledge.

Full Text

D. Cameron, O. Bodenreider, H. Yalamanchili, T. Danh, S. Vallabhaneni, Krishnaprasad Thirunarayan, A. P. Sheth, T. C. Rindflesch, A Graph-Based Recovery and Decomposition of Swanson's Hypothesis using Semantic Predications, Journal of Biomedical Informatics 46(2): 238-251, 2013. http://dx.doi.org/10.1016/j.jbi.2012.09.004 [PMID http://www.ncbi.nlm.nih.gov/pubmed/23026233]


Related Files: