S4: ChemDIS

ChemDIS: in silico analysis of chemical-disease association

ChemDIS, Chun-Wei Tung, OpenTox Asia 2018

Chun-Wei Tung


Kaohsiung Medical University, Taiwan.


Associate Professor


Hazard identification for poorly characterized chemicals has long been a challenge. Given the huge number of chemicals for testing, it is impractical to apply in vivo and in vitro assays to all chemicals. Modern in silico methods for analyzing potential effects induced by chemical exposures provide an efficient and economical way for chemical hazard identification. To help the identification of potential diseases associated with a chemical, we have developed a chemical-disease inference system ChemDIS integrating a large-scale STITCH database of chemical-protein interactions and protein annotation databases including Gene Ontology (GO), KEGG, Reactome, SMPDB and Disease Ontology (DO). Enrichment tools have been developed and integrated into ChemDIS for inferring statistically significant chemical-protein-disease associations. The integration enables the fast identification of chemical-disease association and generation of testable hypotheses for experimental validation. The interacting proteins, functions, pathways and diseases for more than 430,000 unique chemicals can be analyzed in ChemDIS. As a successful example, the effect of maleic acid on neuronal cells has been successfully identified using ChemDIS and subsequently verified by in vitro assays. Several useful functions were also implemented to help the study of potential interaction effects induced by mixture exposure, enrichment analysis of user-supplied differential expression gene, metabolite or miRNA set, and joint analysis of multi-omics data. Web-based application programming interface (API) has been developed for programmatic access to ChemDIS. ChemDIS is expected to be useful for both drug development and hazard identification of environmental chemicals and is publicly accessible at http://cwtung.kmu.edu.tw/chemdis.