Network-based detection of disease modules and potential drug targets in intractable epilepsy

Epilepsy is one of the common nervous system diseases and a complex brain disease that severely damages the life and health of humans. One-third of all epilepsy patients have medically intractable epilepsy (IE), for which anti-epileptic drugs are not effective. Therefore, discovery of potential drug targets is urgent and meaningful for better clinical solutions. Using the IE terms from Medical Subject Headings (MeSH) terminology, we integrated literature-based disease-gene relationships, which were extracted from the CoreMine PubMed search engine system, protein-protein interactions (PPI) and drug-target relationships from heterogeneous data sources, and used the network medicine approach to identify disease modules and detect enriched pathways. The potential drug targets and the underlying mechanisms were confirmed by chemical-protein interaction network and published literatures. Using 23 IE MeSH terms, we manually searched the CoreMine system to obtain 1,400 diseasegene associations, which had 871 distinct genes from the PubMed database. With the help of the PPI database (i.e. String 9), we mapped the genes to the PPI network and utilized the BGL community detection method to find 33 disease-related topological PPI modules that contain 640 proteins and 2,483 links. After that, we used the enrichment analysis method to obtain the PPI modules with pathway and gene ontology enrichment. Finally, we confirmed nine significant PPI modules that are considered as epilepsy disease modules with significant functional signatures. We combined genes with drugs in the DrugBank database to confirm the four proteins, MT-CYB, UQCRB, UQCRC1 and UQCRH, which would be potential drug targets for IE. The results of this study demonstrated that integrated network data sources and network-based approach are useful to understand the molecular mechanism and extract potential drug targets for IE.