From Metabolites to Therapeutics: A Computational Framework for Therapeutic Discovery from Human Urine Metabolites

The emergence of metabolomics as an important concept is attributed to the abundance of metabolites in urine, association of metabolites with physiological and pathophysiological processes, and the intrinsic benefits offered by the technique. This study focuses on finding significant human urine metabolites and studying the possible protein targets of the metabolites by using in-silico approaches. In all, 5603 urine metabolites were retrieved from HMDB and filtered through ADMET filtering process based on oral bioavailability, drug-likeness, and cytochrome P450 inhibition properties. Computer-based search of the selected urine metabolites revealed 709 metabolites with desirable pharmacological properties. Protein targets were found using SEA database. The identified target proteins were analyzed for interaction between proteins using STRING and Cytoscape software, which resulted in the identification of the key target genes such as CYP1A1, CYP1A2, CYP2C9, CYP3A4, AKR1C3, and JUN. Pharmcological property based screened metabolites were subjected to reverse molecular docking process. The protein complex containing phenylalanylphenylalanine-1S2C showed maximum binding affinity compared to all the other complexes. The stability of selected complexes was confirmed using molecular dynamics simulation performed using GROMACS software. These results demonstrate that metabolites obtained from human urine have the possibility of being biomarkers or a possible therapeutic agent in future.