Precision drug repurposing via convergent eQTL-based molecules and pathway targeting independent disease-associated polymorphisms
AffiliationUniv Arizona, Ctr Biomed Informat & Biostat CB2
Univ Arizona, Dept Med COM T
Univ Arizona, Ctr Appl Genet & Genom Med
Univ Arizona, Dept Biosyst Engn
Univ Arizona, BIO5 Inst
Univ Arizona, UA Canc Ctr
Univ Arizona, UAHS
MetadataShow full item record
PublisherWORLD SCIENTIFIC PUBL CO PTE LTD
CitationVitali, F., Berghout, J., Fan, J. W., Li, J., Li, Q., Li, H., & Lussier, Y. A. (2019, January). Precision drug repurposing via convergent eQTL-based molecules and pathway targeting independent disease-associated polymorphisms. In PSB (pp. 308-319).
Rights© 2018 The Authors. Open Access chapter published by World Scientific Publishing Company, distributed under the terms of the Creative Commons Attribution Non-Commercial (CC BY-NC) 4.0 License
Collection InformationThis item from the UA Faculty Publications collection is made available by the University of Arizona with support from the University of Arizona Libraries. If you have questions, please contact us at firstname.lastname@example.org.
AbstractRepurposing existing drugs for new therapeutic indications can improve success rates and streamline development. Use of large-scale biomedical data repositories, including eQTL regulatory relationships and genome-wide disease risk associations, offers opportunities to propose novel indications for drugs targeting common or convergent molecular candidates associated to two or more diseases. This proposed novel computational approach scales across 262 complex diseases, building a multi-partite hierarchical network integrating (i) GWAS-derived SNP-to-disease associations, (ii) eQTL-derived SNP-to-eGene associations incorporating both cis-and trans-relationships from 19 tissues, (iii) protein target-to-drug, and (iv) drug-to-disease indications with (iv) Gene Ontology-based information theoretic semantic (ITS) similarity calculated between protein target functions. Our hypothesis is that if two diseases are associated to a common or functionally similar eGene -and a drug targeting that eGene/protein in one disease exists - the second disease becomes a potential repurposing indication. To explore this, all possible pairs of independently segregating GWAS-derived SNPs were generated, and a statistical network of similarity within each SNP-SNP pair was calculated according to scale-free overrepresentation of convergent biological processes activity in regulated eGenes (ITSeGENE-eGENE) and scale-free overrepresentation of common eGene targets between the two SNPs (ITSSNP-SNP). Significance of ITSSNP-SNP was conservatively estimated using empirical scale-free permutation resampling keeping the node-degree constant for each molecule in each permutation. We identified 26 new drug repurposing indication candidates spanning 89 GWAS diseases, including a potential repurposing of the calcium-channel blocker Verapamil from coronary disease to gout. Predictions from our approach are compared to known drug indications using DrugBank as a gold standard (odds ratio=13.1, p-value=2.49x10(-8)). Because of specific disease-SNPs associations to candidate drug targets, the proposed method provides evidence for future precision drug repositioning to a patient's specific polymorphisms.
NoteOpen access journal
VersionFinal published version
SponsorsUniversity of Arizona Health Sciences CB2; BIO5 Institute; UA Cancer Center; NIH [U01AI122275]