Scholarlink

<p>Nipah virus is a highly lethal zoonotic pathogen for which no approved therapeutics or vaccines are currently available, highlighting the urgent need for new antiviral strategies. In this study, we introduce an integrated computational pipeline for designing and prioritizing peptide inhibitors targeting the HR1 domain of the Nipah virus fusion glycoprotein F, a key factor in viral entry. Starting from the native HR2 sequence, a library of 627 mutant peptides was created, with 200 peptides initially evaluated through protein–protein docking and used to train machine learning regression models based on ESM-2 protein language model embeddings. These trained models were then used to rank the remaining peptides, and the top candidates were further analyzed using flexible docking, MM-GBSA binding free energy calculations, density functional theory (DFT) analysis, and long-timescale (1000 ns) molecular dynamics simulations. Among the peptides tested, Pep_404 consistently showed superior binding affinity, structural stability, favorable electronic properties, and stable interactions with HR1 compared to the native HR2 control. This study showcases a novel end-to-end approach that combines protein language models, machine learning, quantum-chemical analysis, and molecular dynamics simulations to rationally design peptide-based fusion inhibitors against Nipah virus, with potential applications to other class I viral fusion proteins.</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

Inhibition of viral fusion via HR1-binding peptides: a machine learning and molecular dynamics-guided strategy against Nipah virus

  • 期刊论文

  • 发表日期：

    2026年2月17日

查看全文
  • Dharmendrasinh F. Rao,
  • Saumya K. Patel,
  • Himanshu A. Pandya

摘要

Nipah virus is a highly lethal zoonotic pathogen for which no approved therapeutics or vaccines are currently available, highlighting the urgent need for new antiviral strategies. In this study, we introduce an integrated computational pipeline for designing and prioritizing peptide inhibitors targeting the HR1 domain of the Nipah virus fusion glycoprotein F, a key factor in viral entry. Starting from the native HR2 sequence, a library of 627 mutant peptides was created, with 200 peptides initially evaluated through protein–protein docking and used to train machine learning regression models based on ESM-2 protein language model embeddings. These trained models were then used to rank the remaining peptides, and the top candidates were further analyzed using flexible docking, MM-GBSA binding free energy calculations, density functional theory (DFT) analysis, and long-timescale (1000 ns) molecular dynamics simulations. Among the peptides tested, Pep_404 consistently showed superior binding affinity, structural stability, favorable electronic properties, and stable interactions with HR1 compared to the native HR2 control. This study showcases a novel end-to-end approach that combines protein language models, machine learning, quantum-chemical analysis, and molecular dynamics simulations to rationally design peptide-based fusion inhibitors against Nipah virus, with potential applications to other class I viral fusion proteins.