Dataset Information

Machine learning assessment of the binding region as a tool for more efficient computational receptor-ligand docking.

ABSTRACT: We present a combined computational approach to protein-ligand binding, which consists of two steps: (1) a deep neural network is used to locate a binding region on a target protein, and (2) molecular docking of a ligand is performed within the specified region to obtain the best pose using Autodock Vina. Our in-house designed neural network was trained using the PepBDB dataset. Although the training dataset consisted of protein-peptide complexes, we show that the approach is not limited to peptides, but also works remarkably well for a large class of non-peptide ligands. The results are compared with those in which the binding region (first step) was provided by Accluster. In cases where no prior experimental data on the binding region are available, our deep neural network provides a fast and effective alternative to classical software for its localization. Our code is available at https://github.com/mksmd/NNforDocking.

SUBMITTER: Simoncic M

PROVIDER: S-EPMC8903148 | biostudies-literature | 2022 May

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine learning assessment of the binding region as a tool for more efficient computational receptor-ligand docking.

Simončič Matjaž M Lukšič Miha M Druchok Maksym M

Journal of molecular liquids 20220218

We present a combined computational approach to protein-ligand binding, which consists of two steps: (1) a deep neural network is used to locate a binding region on a target protein, and (2) molecular docking of a ligand is performed within the specified region to obtain the best pose using Autodock Vina. Our in-house designed neural network was trained using the PepBDB dataset. Although the training dataset consisted of protein-peptide complexes, we show that the approach is not limited to pept ...[more]

PMID: 35273421

Similar Datasets

Project description:Lung cancer is the most prevalent cause of cancer deaths worldwide. However, its treatment faces a significant hurdle due to the development of resistance. Phytomolecules are an important source of new chemical entities due to their rich chemical diversity. Therefore, a machine learning (ML) model was developed to computationally identify potential inhibitors using a curated data set of 649 phytomolecules with inhibitory activity against lung cancer cell lines. Four distinct ML approaches, including k-nearest neighbor, random forest, support vector machine, and extreme gradient boosting, were used in conjugation with MACCS and Morgan2 fingerprints to generate the models. It was observed that the random forest model developed by using the MACCS fingerprint shows the best performance. To further explore the chemical space and feature importance, k-means clustering, t-SNE analysis, and mean decrease in impurity had been calculated. Simultaneously, ∼400 000 natural products (NPs) retrieved from the COCONUT database were filtered for pharmacokinetic properties and taken for a multistep screening using docking against epidermal growth factor receptor (EGFR) mutant, a therapeutic drug target of lung cancer. Thereafter, the best-performing random forest model was used to predict the antilung cancer potential of the NPs having binding affinity better than the cocrystal ligand. This allowed the identification of 205 potential inhibitors, wherein the molecules with an indolocarbazole scaffold were enriched in top-scoring molecules. The top three indolocarbazole molecules with the lowest binding energy were further evaluated through 100 ns molecular dynamics (MD) simulations, which suggested that these molecules are strong binders. Also, structural similarity analysis against known drugs revealed that these NPs are similar to staurosporine, which demonstrates potent and selective activity against EGFR mutants. Thereby, the consensus analysis employing ML, molecular docking, and dynamics revealed that the molecules having an indolocarbazole scaffold are the most promising NPs that can act as potential inhibitors against lung cancer.

Project description:BackgroundCOVID-19 is an infectious disease caused by SARS-CoV-2, a close relative of SARS-CoV. Several studies have searched for COVID-19 therapies. The topics of these works ranged from vaccine discovery to natural products targeting the SARS-CoV-2 main protease (Mpro), a potential therapeutic target due to its essential role in replication and conserved sequences. However, published research on this target is limited, presenting an opportunity for drug discovery and development.MethodThis study aims to repurpose 10692 drugs in DrugBank by using ligand-based virtual screening (LBVS) machine learning (ML) with Konstanz Information Miner (KNIME) to seek potential therapeutics based on Mpro inhibitors. The top candidate compounds, the native ligand (GC-376) of the Mpro inhibitor, and the positive control boceprevir were then subjected to absorption, distribution, metabolism, excretion, and toxicity (ADMET) characterization, drug-likeness prediction, and molecular docking (MD). Protein-protein interaction (PPI) network analysis was added to provide accurate information about the Mpro regulatory network.ResultsThis study identified 3,166 compound candidates inhibiting Mpro. The random forest (RF) molecular access system ML model provided the highest confidence score of 0.95 (bromo-7-nitroindazole) and identified the top 22 candidate compounds. Subjecting the 22 candidate compounds, the native ligand GC-376, and boceprevir to further ADMET property characterization and drug-likeness predictions revealed that one compound had two violations of Lipinski's rule. Additional MD results showed that only five compounds had more negative binding energies than the native ligand (- 12.25 kcal/mol). Among these compounds, CCX-140 exhibited the lowest score of - 13.64 kcal/mol. Through literature analysis, six compound classes with potential activity for Mpro were discovered. They included benzopyrazole, azole, pyrazolopyrimidine, carboxylic acids and derivatives, benzene and substituted derivatives, and diazine. Four pathologies were also discovered on the basis of the Mpro PPI network.ConclusionResults demonstrated the efficiency of LBVS combined with MD. This combined strategy provided positive evidence showing that the top screened drugs, including CCX-140, which had the lowest MD score, can be reasonably advanced to the in vitro phase. This combined method may accelerate the discovery of therapies for novel or orphan diseases from existing drugs.

Dataset Information

Machine learning assessment of the binding region as a tool for more efficient computational receptor-ligand docking.

Publications

Machine learning assessment of the binding region as a tool for more efficient computational receptor-ligand docking.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets