Identifying Phytochemicals from Biomedical Literature Utilizing Semantic Knowledge Sources

Chemicals derived from plants (phytochemicals) are major concepts of interest in the study of medicinal plants. To date, efforts to catalogue and organize phytochemical knowledge have resorted to manual approaches. This study explored the potential to leverage publicly accessible semantic knowledge sources for identifying possible phytochemicals. Within the context of this feasibility study, putative phytochemicals were identified for more than 4,000 plants from the Medical Subject Headings Supplementary Concept Records and the Semantic MEDLINE Database. An examination of phytochemicals identified for five selected plant species using the method developed here reveals that there is a disparity in electronically catalogued phytochemical knowledge compared to information from Dr. Duke's Phytochemical and Ethnobotanical Databases maintained by the United States Department of Agriculture. The results therefore suggest that semantic knowledge sources for biomedicine can be utilized as a source for identifying potential phytochemicals and thus contribute to the overall curation of plant phytochemical knowledge.