Virtual Screening on Indonesian Herbal Compounds as COVID-19 Supportive Therapy: Machine Learning and Pharmacophore Modelling Approaches

Abstract

Background: The latest development of COVID-19 spread in Indonesia has reached 311,176 cases, with 11,374 patients died, updated on October 6, 2020. Unfortunately, these numbers continue to overgrow, and no drug has yet been approved for effective treatment. This study aims to determine the potential candidate compounds in Indonesian herbal medicine as a COVID-19 supportive therapy using a machine learning and pharmacophore modelling approach.Methods: For the machine learning approach, we used three classification methods that have different ways in decision making, such as Support Vector Machine (SVM), Multilayer Perceptron (MLP), and Random Forest (RF). Moreover, for the pharmacophore modelling approach, we performed a structure-based method on the 3D structure of the main protease SARS-CoV-2 (3CLPro) and used the SARS, MERS, and SARS-CoV-2 repurposing drugs known from the literature as data sets on the ligand-based method. Finally, we used molecular docking to analyse the interactions between the 3CLpro protein (main protease) and 14 hit compounds from the Indonesian Herbal Database (HerbalDB) and Lopinavir as a positive control.Results: The machine learning approach with SVM, RF, and MLP methods and pharmacophore modelling approach were used for screening in herbal compounds obtained from HerbalDB. Based on the screening on HerbalDB using these two prediction approaches, we got 14 hit compounds. We then performed molecular docking to determine the interaction of these compounds with the main protease SARS-CoV-2 as an inhibiting agent. From the molecular docking analysis, it was found that six potential compounds as the main proteases of the SARS-CoV-2 inhibitor, i.e. Hesperidin, Kaempferol-3,4'-di-O-methyl ether (Ermanin); Myricetin-3-glucoside, Peonidine 3-(4’-arabinosylglucoside); Quercetin 3-(2G-rhamnosylrutinoside); and Rhamnetin 3-mannosyl-(1-2)-alloside.Conclusions: We used layered virtual screening with machine learning and pharmacophore modelling approaches that could provide more objective and optimal virtual screening and avoid subjective decision making on research results. Herbal compounds from various plants have potential as antiviral candidates for SARS-CoV-2. Based on our research and literature study, one of Indonesia's potential commodity crops is Psidium guajava (guava), and people can use it directly as a preventive effort.

Focusing on development of novel sampling approaches and alternative therapies for COVID-19: Are they still useful in an era after the pandemic?

“…Studies in Mexico even orally administered chlorine dioxide (ClO 2 ) in a low dose to prevent/treat COVID-19-related symptoms in relatives living with COVID-19 patients (7,8). iii) Alleviation of symptoms using natural products: This scenario includes a number of natural products such as Indonesian herbal compounds (9), Iran herbal medicine (10), propolis and honey (11), Indian herbs (12), essential oils (13), and traditional Chinese medicine (TCM) (14)(15)(16). Although all of these studies contend that these natural products "are efficacious and safe", these products are not amenable to becoming generally accepted therapeutics worldwide due to regional and cultural characteristics.…”

Machine Learning for the Prediction of Antiviral Compounds Targeting Avian Influenza A/H9N2 Viral Proteins

“…We prefer to use the oversampling method over the undersampling method, because in the previous studies with cases of unbalanced data and the same type of data, using oversampling to balance the data is more effective. The result is that the ensemble method model is more robust because oversampling increases the probability of the classes needed to construct the tree during the node splitting process [40]. The oversampling method duplicates the data in the minor class so that the size is the same as the data in the major class.…”

Active constituents and Molecular Analysis of Psidium guajava Against Multiple Protein of SARS-CoV-2

“…(2020), found that Psidium guajava is potential as a candidate of SARS-CoV-2 therapeutic agent. By using combination studies of machine learning, pharmacophore modelling and molecular docking, some of Psidium guajava compounds such as myricetin, quercetin, luteolin, kaempferol, isorhamnetin [6], and Hesperidin [7] shown good activity for 3CL Pro inhibition [8].…”

“…8.72 kcal/mol, RMSD value 3.96 Å and inhibition constant value 408.59 nM. The optimum grid center is x= -9.768; y = 11.424; z = 68.935 with 0.375 Å spacing for default setting.…”