Prioritizing Pain-Associated Targets with Machine Learning

• DOI: 10.1021/acs.biochem.0c00930
• Publication Date: 2021-02-20T07:14:20Z
• Authors: Minji Jeon; Kathleen M. Jagodnik; Eryk Kropiwnicki; Daniel J. Stein; Avi Ma’ayan
• Keywords: drug target families; Information Systems not elsewhere classified; prioritize targets; Biophysics; Pain; novel pain targets; modulating pain; Models, Biological; Machine Learning; 03 medical and health sciences; GPCR; Drug Delivery Systems; GPR 132; Prioritizing Pain-Associated; GPR 109B; Drug Discovery; Genetics; Humans; ensemble model; protein kinases; Analgesics; 0303 health sciences; AUROC; ML; novel G-protein-coupled receptors; GABA-related cell; gene-pain association predictions c.; Drug Design; 17 categories; gene ontology; Biological Sciences not elsewhere classified; novel targets

While hundreds of genes have been associated with pain, much the molecular mechanisms pain remain unknown. As a result, current analgesics are limited to few clinically validated targets. Here, we trained machine learning (ML) ensemble model predict new targets for 17 categories pain. The utilizes features from transcriptomics, proteomics, and gene ontology prioritize modulating We focused on identifying novel G-protein-coupled receptors (GPCRs), ion channels, protein kinases because these proteins represent most successful drug target families. performance is 0.839 average based AUROC, while predictions arthritis had highest accuracy (AUROC = 0.929). predicts pain; example, GPR132 GPR109B highly ranked GPCRs rheumatoid arthritis. Overall, gene-pain association cluster into three groups that enriched cytokine, calcium, GABA-related cell signaling pathways. These can serve as foundation future experimental exploration advance development safer more effective analgesics.