Abstract:Background: Recently, the metabolite-likeness of the drug space has emerged and has opened a new possibility for exploring human metabolite-like candidates in drug discovery. However, the applicability of metabolite-likeness in drug discovery has been largely unexplored. Moreover, there are no reports on its applications for the repositioning of drugs to possible enzyme modulators, although enzyme-drug relations could be directly inferred from the similarity relationships between enzyme's metabolites and drugs. Methods: We constructed a drug-metabolite structural similarity matrix, which contains 1,861 FDA-approved drugs and 1,110 human intermediary metabolites scored with the Tanimoto similarity. To verify the metabolite-likeness measure for drug repositioning, we analyzed 17 known antimetabolite drugs that resemble the innate metabolites of their eleven target enzymes as the gold standard positives. Highly scored drugs were selected as possible modulators of enzymes for their corresponding metabolites. Then, we assessed the performance of metabolite-likeness with a receiver operating characteristic analysis and compared it with other drug-target prediction methods. We set the similarity threshold for drug repositioning candidates of new enzyme modulators based on maximization of the Youden's index. We also carried out literature surveys for supporting the drug repositioning results based on the metabolite-likeness. Results: In this paper, we applied metabolite-likeness to repurpose FDA-approved drugs to disease-associated enzyme modulators that resemble human innate metabolites. All antimetabolite drugs were mapped with their known 11 target enzymes with statistically significant similarity values to the corresponding metabolites. The comparison with other drug-target prediction methods showed the higher performance of metabolite-likeness for predicting enzyme modulators. After that, the drugs scored higher than similarity score of 0.654 were selected as possible modulators of enzymes for their corresponding metabolites. In addition, we showed that drug repositioning results of 10 enzymes were concordant with the literature evidence. Conclusions: This study introduced a method to predict the repositioning of known drugs to possible modulators of disease associated enzymes using human metabolite-likeness. We demonstrated that this approach works correctly with known antimetabolite drugs and showed that the proposed method has better performance compared to other drug target prediction methods in terms of enzyme modulators prediction. This study as a proof-of-concept showed how to apply metabolite-likeness to drug repositioning as well as potential in further expansion as we acquire more disease associated metabolite-target protein relations.

Deep learning large-scale drug discovery and repurposing

Deep learning identifies explainable reasoning paths of mechanism of action for drug repurposing from multilayer biological network

Deep learning-driven neuromorphogenesis screenings identify repurposable drugs for mitochondrial disease

deepDR: a network-based deep learning approach to in silico drug repositioning

De Novo Generation and Identification of Novel Compounds with Drug Efficacy Based on Machine Learning

Large-Scale Off-Target Identification Using Fast and Accurate Dual Regularized One-Class Collaborative Filtering and Its Application to Drug Repurposing.

Morphological Profiling for Drug Discovery in the Era of Deep Learning

Artificial Intelligence/Machine Learning-Driven Small Molecule Repurposing via Off-Target Prediction and Transcriptomics

Drug Repurposing Pipeline for Epilepsy: Integrating Deep Learning and Structure-based Approaches

Drug repositioning for enzyme modulator based on human metabolite-likeness

Open MoA: revealing the mechanism of action (MoA) based on network topology and hierarchy

In Silico drug repurposing pipeline using deep learning and structure based approaches in epilepsy

DrugReAlign: a multisource prompt framework for drug repurposing based on large language models

DeepDrug: An Expert-led Domain-specific AI-Driven Drug-Repurposing Mechanism for Selecting the Lead Combination of Drugs for Alzheimer's Disease

DeepREAL: A Deep Learning Powered Multi-scale Modeling Framework Towards Predicting Out-of-distribution Receptor Activity of Ligand Binding

An Integrated Deep Learning and Molecular Dynamics Simulation-Based Screening Pipeline Identifies Inhibitors of a New Cancer Drug Target TIPE2

Nanoscale Monitoring of Mitochondria and Lysosome Interactions for Drug Screening and Discovery

Deep Learning Based Drug Metabolites Prediction

DLDTI: a learning-based framework for drug-target interaction identification using neural networks and network representation

DrugRepPT: a Deep Pre-Training and Fine-Tuning Framework for Drug Repositioning Based on Drug's Expression Perturbation and Treatment Effectiveness.

Deep learning in target prediction and drug repositioning: Recent advances and challenges