Semantic data integration to biological relationship among chemicals, diseases, and differential expressed genes

Systems approaches are showing early promise in helping bridge the gap between pathophysiological processes and their molecular determinants. In toxicology, microarray technology leads rapid screening of DEGs (differential expressed gene) from various kinds of chemical exposes. Using toxicogenomics for the risk assessment, various and heterogeneous data are contributed to each step, such as genome sequence, genotype, gene expression, phenotype, disease information, etc. To derive actual roles of the DEGs, it is essentially required to construct interactions among DEGs and to link the known information of diseases. Proper data model is essential and critical component to build information system for risk assessment. Our study suggests a semantic modeling strategy to organize heterogeneous data types and introduces techniques and concepts (such as ontologies, semantic objects, typed relationships, contexts, graphs, and information layers) that are used to represent complex biomedical networks. We depict reconstruction of semantic relationship among chemicals, diseases, and DEGs in public available data. In this work, user’s experiment results can be easily uploaded and bound to the current data network. This feature provides to maintain user’s specific interactions from their interesting DEGs to publicly available disease and chemical data. The program was built upon DjangoWeb framework in Python language and commercial text-mining engine, MedScan, was employed. Example analysis was completed for evaluation of the system and presented in this paper. We are expecting that this work provides rapid way to build custom-driven toxico-knowledgebase by integrating customers internal documents and public data.