key: cord-0952925-c9rfyv87
authors: Rajput, Akanksha; Thakur, Anamika; Mukhopadhyay, Adhip; Kamboj, Sakshi; Rastogi, Amber; Gautam, Sakshi; Jassal, Harvinder; Kumar, Manoj
title: Prediction of repurposed drugs for Coronaviruses using artificial intelligence and machine learning
date: 2021-05-24
journal: Comput Struct Biotechnol J
DOI: 10.1016/j.csbj.2021.05.037
sha: c5da044baa349356f6fd03bcfa199fac3d213195
doc_id: 952925
cord_uid: c9rfyv87

The world is facing the COVID-19 pandemic caused by Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2). Likewise, other viruses of the Coronaviridae family were responsible for causing epidemics earlier. To tackle these viruses, there is a lack of approved antiviral drugs. Therefore, we have developed robust computational methods to predict the repurposed drugs using machine learning techniques namely Support Vector Machine, Random Forest, k-Nearest Neighbour, Artificial Neural Network, and Deep Learning. We used the experimentally validated drugs/chemicals with anti-corona activity and their inhibition efficiencies (IC(50)/EC(50)) from ‘DrugRepV’ repository. The unique entries of SARS-CoV-2 (142), SARS (221), MERS (123), and overall Coronaviruses (414) were subdivided into the training/testing and independent validation datasets, followed by the extraction of chemical/structural descriptors and fingerprints (17968). The highly relevant features were filtered using the recursive feature selection algorithm. The selected chemical descriptors were used to develop prediction models with Pearson’s correlation coefficients ranging from 0.60-0.90 on training/testing. The robustness of the predictive models was further ensured using external independent validation datasets, decoy datasets, applicability domain, and chemical analyses. The developed models were used to predict promising repurposed drug candidates against coronaviruses after scanning the DrugBank. Top predicted molecules for SARS-CoV-2 were further validated by molecular docking against the spike protein complex with ACE receptor. We found potential repurposed drugs namely, Verteporfin, Alatrofloxacin, Metergoline, Rescinnamine, Leuprolide, and Telotristat ethyl with high binding affinity. These computational methods would assist in antiviral drug discovery against SARS-CoV-2 and other Coronaviruses.

The 21 st century has experienced three novel coronavirus (CoV) pandemics caused by the Severe identification of the epitopes for the CoVs named 'CoronaVR' [31] . The input anti-CoVs data in 87 the current study was taken from our recently published comprehensive database of the 88 experimentally validated repurposed drug database named 'DrugRepV' [32] . In the current study, 89 we have identified repurposed drug candidates (against SARS-CoV-2, SARS, and MERS) using 120 For SARS, various prediction models were developed using the MLTS like SVM, RF, kNN, and 121 ANN. The performance of the training/testing dataset with 198 datasets was calculated using the 122 10-fold cross-validation ( Table 1 ). The prediction model developed using the training/testing Table S2 ).

The prediction models were also developed for the MERS using 10-fold cross-validation on 130 training/testing and independent validation datasets ( Table 1) Table S2 ).

The SARS-CoV-2 dataset was subdivided into 127 training/testing and 15 independent validation 138 dataset ( Table 1) Table S2 ).

The Overall CoVs include unique entries from the SARS, MERS, and SARS-CoV-2 datasets. The 145 overall entries were split into the training/testing and independent validation datasets with 372 and 146 42 entries via the randomization approach available in SciKit library ( 154 The applicability domain was calculated between the leverage and the standardized residuals 155 among the best performing SVM models. All the models of SVM on the SARS, SARS-CoV-2, The best performing SVM predictors were used to identify the repurposed drug candidates against 184 SARS, MERS, and SARS-CoV-2 (Figure 3c, Supplementary Figures S2-S4 The molecular docking technique is highly beneficial for understanding the protein-ligand Table 4 represents the interacting residues, interacting domain of the protein, type of 225 interactions, as well as bond length of the 06 ligands mentioned above. Figure S5) . The detail of the top hits predicted from our pipeline, 233 DrugBank ids, drug name, primary indication, status, and references is provided in Table 3 . 234 From predicted drugs for the SARS-CoV-2, 17 drugs are already found in clinical trials like (Table 3) . However, some drugs also validated through in vivo studies, e.g., Nilotinib showing 242 inhibition win Vero-E6 cells and Calu-3 cells with EC 50 of 1.44 μM and 3.06 μM, respectively, 243 while Bosutinib shows EC 50 of 2.45 ± 0.12 μM for SARS-CoV-2. Thus, this analysis demonstrates 244 the robustness of our prediction algorithm, which further suggests that there is a higher probability 245 that the predicted drugs will show promising results against the SARS-CoV-2. Since drug development is a very complex and time-consuming process, from the start of the 317 SARS-CoV-2 pandemic, several research groups have been trying to identify efficient repurposed 318 drug candidates via computational, in vitro, and in vivo studies. So our developed computational 319 predictive models were used to identify the repurposed drug candidates from the "approved" drug 320 category of the DrugBank database. Further, we checked the predicted repurposed drug candidates 321 using our pipeline, which have been already validated in the literature. Interestingly, we found that The datasets used in the analysis are available as Supplementary Tables S5-S8 . 376 The overall methodology is described in Figure 6 . The following steps have been used: In general, its main focus is minimizing the error, maximizing the margin by individualizing the 431 hyperplane, such that some proportion of the error is being tolerated. It was customized by using Repurposing of the drugs against the SARS-CoV-2, SARS, and MERS was done using our 541 developed predicted models. We predicted the repurposed drugs using the best performing SVM 542 models in all three categories. For repurposing the drug categories the "Approved" category of the 543 drugs was downloaded from the DrugBank repository [65] . The descriptors and fingerprints of all 544 the 2468 approved drugs were calculated using the PaDel software. Further, the descriptors of the 545 approved drugs were used to predict the highly efficient drugs against all three categories of 546 viruses.

The AutoDock tool (ADT) was used to customize the ligand and Protein [66] . Further, their Code availability 612 The Python code used in study is provided on GitHub (https://github.com/manojk-613 imtech/antiCorona).

Conflict of interest 616 The authors declare that the research was conducted in the absence of any commercial or financial 617 relationships that could be construed as a potential conflict of interest. 

SARS and MERS: recent 670 insights into emerging coronaviruses

Cryo-electron tomography of mouse hepatitis virus: Insights into the structure of the 673 coronavirion

Supramolecular architecture of severe acute respiratory syndrome coronavirus revealed by 676 electron cryomicroscopy

Discovery 678 of seven novel Mammalian and avian coronaviruses in the genus deltacoronavirus supports 679 bat coronaviruses as the gene source of alphacoronavirus and betacoronavirus and avian 680 coronaviruses as the gene source of gammacoronavirus and deltacoronavirus

Genomic 683 characterization of the 2019 novel human-pathogenic coronavirus isolated from a patient with 684 atypical pneumonia after visiting Wuhan

Genomic characterisation and 686 epidemiology of 2019 novel coronavirus: implications for virus origins and receptor binding

The molecular biology of coronaviruses

Efficient assembly and release of SARS coronavirus-like 691 particles by a heterologous expression system

Antigenicity of the SARS-CoV-2 Spike Glycoprotein

Pre-fusion structure of a human coronavirus spike protein

Coronavirus in human diseases: 698 Mechanisms and advances in clinical treatment

New understanding of the damage 701 of SARS-CoV-2 infection outside the respiratory system

Comparative 704 therapeutic efficacy of remdesivir and combination lopinavir, ritonavir, and interferon beta 705 against MERS-CoV

Remdesivir and chloroquine 707 effectively inhibit the recently emerged novel coronavirus (2019-nCoV) in vitro

Chloroquine is a potent inhibitor of SARS coronavirus infection and spread

Screening of an FDA-approved compound library identifies four 714 small-molecule inhibitors of Middle East respiratory syndrome coronavirus replication in cell 715 culture

Chloroquine for the 2019 novel coronavirus 717 SARS-CoV-2

Hydroxychloroquine, a less 719 toxic derivative of chloroquine, is effective in inhibiting SARS-CoV-2 infection in vitro

SARS-CoV-2: Recent Reports on 722 Antiviral Therapies Based on Lopinavir/Ritonavir, Darunavir/Umifenovir

Favipiravir and other Drugs for the Treatment of the New 724

lopinavir/ritonavir in the treatment of SARS: initial virological and clinical findings

Combination therapy with

lopinavir/ritonavir, ribavirin and interferon-α for Middle East respiratory syndrome

Treatment of Middle East Respiratory Syndrome with a combination of lopinavir-ritonavir 733 and interferon-β1b (MIRACLE trial): study protocol for a randomized controlled trial

Triple 736 combination of interferon beta-1b, lopinavir-ritonavir, and ribavirin in the treatment of 737 patients admitted to hospital with COVID-19: an open-label, randomised, phase 2 trial

Lopinavir/ritonavir for the treatment of SARS, 740 MERS and COVID-19: a systematic review

Molecular docking and dynamic simulations for antiviral compounds against SARS-CoV-2: 743 A computational study

Identification 745 of potential inhibitors against SARS-CoV-2 by targeting proteins responsible for envelope 746 formation and virion assembly using docking based virtual screening, and pharmacokinetics 747 approaches

AVPpred: collection and prediction of highly 749 effective antiviral peptides

AVP-IC50 Pred: Multiple machine learning 751 techniques-based prediction of peptide antiviral activity in terms of half maximal inhibitory 752 concentration (IC50)

HIVprotI: an integrated web based 754 platform for prediction and design of HIV proteins inhibitors

Anti-flavi: A Web Platform to Predict Inhibitors of 756

CoronaVR: A Computational Resource and Analysis of Epitopes and Therapeutics for Severe

DrugRepV: a compendium of 762 repurposed drugs and chemicals targeting epidemic and pandemic viruses

The SARS, MERS 765 and novel coronavirus (COVID-19) epidemics, the newest and biggest global health threats: 766 what lessons have we learned?

Protoporphyrin IX and 768 verteporfin potently inhibit SARS-CoV-2 infection and in a mouse model expressing human 769 ACE2

Mechanisms of action of photodynamic therapy with 771 verteporfin for the treatment of age-related macular degeneration

Multifunctional theranostic Pluronic mixed micelles improve targeted photoactivity of 775 Verteporfin in cancer cells

Neurocovid: Pharmacological Recommendations for Delirium Associated With

Prediction of Novel Inhibitors of the Main Protease (M-pro) of

SARS-CoV-2 through Consensus Docking and Drug Reposition

Thrombin Inhibition by Argatroban: Potential 784 Therapeutic Benefits in COVID-19. Cardiovasc Drugs Ther

Targeting the SARS-CoV-2 Main Protease to Repurpose Drugs for COVID-19

Computational Identification of Inhibitors Using 790 QSAR Approach Against Nipah Virus

Computational evidence for nitro derivatives of quinoline and quinoline N-oxide as low-793 cost alternative for the treatment of SARS-CoV-2 infection

In-silico drug 795 repurposing study: Amprenavir, enalaprilat, and plerixafor, potential drugs for destabilizing 796 the SARS-CoV-2 S-protein-angiotensin-converting enzyme 2 complex

High-throughput 799 virtual screening of drug databanks for potential inhibitors of SARS-CoV-2 spike 800 glycoprotein

Repurposing of FDA-Approved Toremifene to Treat

COVID-19 by Blocking the Spike Glycoprotein and NSP14 of SARS-CoV-2

Drug design and repurposing with DockThor-VS web server focusing on SARS-CoV-2 808 therapeutic targets and their non-synonym variants

Network Pharmacology and 810 bioinformatics analyses identify intersection genes of niacin and COVID-19 as potential 811 therapeutic targets

Exploiting cheminformatic 813 and machine learning to navigate the available chemical space of potential small molecule 814 inhibitors of SARS-CoV-2

Predicting commercially available 816 antiviral drugs that may act on the novel coronavirus (SARS-CoV-2) through a drug-target 817 interaction deep learning model

Network-based drug 819 repurposing for novel coronavirus 2019-nCoV/SARS-CoV-2

PaDEL-descriptor: an open source software to calculate molecular 821 descriptors and fingerprints

Quantitative structure-activity relationship 823 methods: perspectives on drug discovery and toxicology

AVCpred: an integrated web server for prediction 826 and design of antiviral compounds

Open Babel: An open chemical toolbox

Support Vector Machines. Encyclopedia of Algorithms

Machine Learning with Scikit-Learn and Tensorflow: Deep Learning 832 with Python (Random Forests, Decision Trees, and Neural Networks). Independently 833 Published

Pardalos PM. k-Nearest Neighbor Classification. Data 835 Mining in Agriculture

Artificial Neural Network. Encyclopedia of Sciences and Religions

Concepts, Tools, and Techniques to Build Intelligent Systems

A support vector machine-841 recursive feature elimination feature selection method based on artificial contrast variables 842 and mutual information

Recursive feature elimination 844 for brain tumor classification using desorption electrospray ionization mass spectrometry 845 imaging

RADER: a RApid DEcoy Retriever to 847 facilitate decoy based assessment of virtual screening

ChemMine tools: an online service for analyzing 849 and clustering small molecules

DrugBank 851 5.0: a major update to the DrugBank database

AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility

AutoDock Vina: improving the speed and accuracy of docking 857 with a new scoring function, efficient optimization, and multithreading

Structural and Functional 860 Basis of SARS-CoV-2 Entry by Using Human ACE2

Virtual screening based on molecular docking of possible inhibitors of 863

Covid-19 main protease

Akanksha Rajput: Methodology, Data Curation, Software, Investigation, Validation, Formal analysis

Writing -Original Draft, Writing -review & editing

Curation, Validation, Formal analysis, Visualization, Writing -Original Draft

Validation, Formal analysis, Writing -Original Draft. Sakshi Kamboj: Validation, Formal analysis

Amber Rastogi: Validation, Formal analysis, Writing 916 -Original Draft. Sakshi Gautam: Validation, Writing -Original Draft, Writing -review & editing

Harvinder Jassal: Visualization. Manoj Kumar: Conceptualization, Supervision, Formal analysis, 918 Funding acquisition, Project administration, Writing -original draft, Writing -Original Draft