Scholarlink

<p>Dengue virus serotype-2 (DENV-2) significantly contributes to the global dengue burden, with a rising incidence in endemic regions like India and Southeast Asia. The urgency for effective antiviral therapeutics is heightened by the absence of approved antiviral drugs and the predominance of DENV-2 in severe disease manifestations. This study employed advanced computational techniques, integrating molecular docking, molecular dynamics (MD) simulations, and machine learning models, to identify inhibitors targeting key DENV-2 proteins: E-protein, NS1, NS2B-NS3 protease, and NS5 (RNA-dependent RNA polymerase and methyltransferase domains). A curated library of 189 bioactive compounds derived from natural sources underwent drug-likeliness and ADME profiling, ensuring compliance with Lipinski’s rule of five and additional pharmacokinetic filters. Molecular docking revealed campesterol (E-protein), naringin (NS1), balsacone C (NS2B-NS3 protease), silibinin (NS5 methyltransferase), and ursolic acid (NS5 RNA polymerase) as the top-scoring inhibitors, exhibiting binding affinities ranging from -7.2 to -10.1&#xa0;kcal/mol, surpassing doxycycline, the reference control. MD simulations (100&#xa0;ns) validated the stability and conformational flexibility of the selected protein–ligand complexes, with analyses of RMSD, RMSF, hydrogen bonding, solvent-accessible surface area (SASA), and radius of gyration (Rg). Principal Component Analysis (PCA) confirmed ligand-induced conformational stabilization. Additionally, a hybrid ML/DL-based docking score prediction model (FNN-KNN-RF) was developed, and deployed at <a href="https://github.com/Vishu1197/docking-predictor">https://github.com/Vishu1197/docking-predictor</a>, refining ligand selection by cross-validating docking results. This study provides a robust computational framework for dengue drug discovery, bridging in-silico predictions with experimental validation, and sets the foundation for future antiviral therapeutic development targeting DENV and other viral pathogens.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Unveiling natural antiviral agents against dengue virus: a hybrid machine learning and molecular dynamics approach

  • 期刊论文

  • 发表日期：

    2025年11月28日

查看全文
  • Vishal Chanda,
  • Padmini Topinar Hanumantharayudu,
  • Vivek Keshri,
  • Agnik Haldar,
  • Yuvashree Muralidaran,
  • Prabhakar Mishra

摘要

Dengue virus serotype-2 (DENV-2) significantly contributes to the global dengue burden, with a rising incidence in endemic regions like India and Southeast Asia. The urgency for effective antiviral therapeutics is heightened by the absence of approved antiviral drugs and the predominance of DENV-2 in severe disease manifestations. This study employed advanced computational techniques, integrating molecular docking, molecular dynamics (MD) simulations, and machine learning models, to identify inhibitors targeting key DENV-2 proteins: E-protein, NS1, NS2B-NS3 protease, and NS5 (RNA-dependent RNA polymerase and methyltransferase domains). A curated library of 189 bioactive compounds derived from natural sources underwent drug-likeliness and ADME profiling, ensuring compliance with Lipinski’s rule of five and additional pharmacokinetic filters. Molecular docking revealed campesterol (E-protein), naringin (NS1), balsacone C (NS2B-NS3 protease), silibinin (NS5 methyltransferase), and ursolic acid (NS5 RNA polymerase) as the top-scoring inhibitors, exhibiting binding affinities ranging from -7.2 to -10.1 kcal/mol, surpassing doxycycline, the reference control. MD simulations (100 ns) validated the stability and conformational flexibility of the selected protein–ligand complexes, with analyses of RMSD, RMSF, hydrogen bonding, solvent-accessible surface area (SASA), and radius of gyration (Rg). Principal Component Analysis (PCA) confirmed ligand-induced conformational stabilization. Additionally, a hybrid ML/DL-based docking score prediction model (FNN-KNN-RF) was developed, and deployed at https://github.com/Vishu1197/docking-predictor, refining ligand selection by cross-validating docking results. This study provides a robust computational framework for dengue drug discovery, bridging in-silico predictions with experimental validation, and sets the foundation for future antiviral therapeutic development targeting DENV and other viral pathogens.