Pharmacol Res. 2020 Sep 17:105203. doi: 10.1016/j.phrs.2020.105203. Online ahead of print.
Even though many genetic risk loci for human diseases have been identified and comprehensively cataloged, strategies to guide clinical research by integrating the extensive results of genetic studies and biological resources are still limited. Moreover, integrative analyses that provide novel insights into disease biology are expected to be especially useful for drug discovery. Herein, we use text mining of genetic studies on colorectal cancer (CRC) and assign biological annotations to
identified risk genes in order to discover novel drug targets and potential drugs for repurposing. Risk genes for CRC were obtained from PubMed text mining, and for each gene, six functional and bioinformatic annotations were analyzed. The annotations include missense mutations, expression quantitative trait loci (eQTL), molecular pathway analyses, protein-protein interactions (PPIs), genetic overlap with knockout mouse phenotypes, and primary immunodeficiency. We then prioritized biological risk candidate genes according to a scoring system for the six functional annotations. Each functional annotation was assigned one point, and those genes with a score ≥2 were designated "biological CRC risk genes". Using this method, we revealed 82 biological CRC risk genes, which mapped to 128 genes in an expanded PPI network. Further utilizing DrugBank and the Therapeutic Target Database (TTD), we found 21 genes in our list that are targeted by 166 candidate drugs. Based on data from ClinicalTrials.gov, we found our list contains four known target genes with six drugs approved for CRC treatment, as well as three known target genes with nine drugs under preclinical investigation for CRC. Additionally, 12 genes are targeted by 32 drugs approved for other indications, which can possibly be repurposed for CRC treatment. Finally, our analysis from Connectivity Map (CMap) showed that 18 of the 41 drugs under clinical and preclinical investigation have high potential to be useful for CRC.