Scholarlink

<p><i>Campylobacter jejuni</i> is a leading cause of bacterial gastroenteritis worldwide, and its growing resistance to antibiotics highlights the urgent need for novel therapeutic agents. This study presents an AI-driven in silico pipeline to identify marine derived bioactive compounds with potential antibacterial activity against <i>C. jejuni</i>. A total of 500 marine natural products were screened, and 58 compounds were selected after filtering using Lipinski’s Rule of Five Veber’s criteria, and ADMET profiling. Fascaplysine, Auranthine, and Ascididemin emerged as the top lead candidates based on favorable pharmacokinetic and safety profiles. Network pharmacology identified six key hub genes TLR4, TNF, MMP3, CYP27B1, ACE, and STAT1 associated with host immune modulation and <i>C. jejuni</i> pathogenesis. These targets are enriched in immune signaling pathways including Toll-like receptor signaling, Necroptosis, and Inflammatory Bowel Disease. Molecular docking using AutoDock 4.2.6 revealed strong binding affinities, with the Fascaplysine-STAT1 complex exhibiting the highest affinity (-9.96&#xa0;kcal/mol; K<sub>i</sub> = 49.53 nM). All three marine lead compounds demonstrated superior binding affinities compared to ciprofloxacin, the clinically approved first-line antibiotic, at all targets for which they were the top-ranked hits; the difference was most pronounced at STAT1, where Fascaplysine (-9.96&#xa0;kcal/mol) substantially outperformed ciprofloxacin (-6.69&#xa0;kcal/mol), a difference of 3.27&#xa0;kcal/mol. A Random Forest regression model trained on dockingmderived features and molecular descriptors achieved an R² of approximately 0.90 outperforming XGBoost, AdaBoost and KNN in binding affinity prediction. A comparative analysis between AutoDock-derived and Random Forest-predicted binding affinities across 30 protein ligand pairs demonstrated strong concordance (mean |ΔAffinity| = 0.66&#xa0;kcal/mol), confirming the predictive accuracy of the model. The model was integrated into a publicly accessible web application enabling real-time binding affinity prediction using PubChem compound descriptors. Off-target screening via SwissTargetPrediction indicated low predicted cross-reactivity with non-target human proteins for all three lead compounds. Together these findings highlight marine-derived alkaloids as promising antibacterial candidates and provide an AI-enhanced computational framework to accelerate early-stage drug discovery against antibiotic-resistant <i>C. jejuni</i>.</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

AI-driven in silico discovery of marine bioactive compounds targeting campylobacter jejuni: network pharmacology, molecular docking, and binding affinity prediction

  • 期刊论文

  • 发表日期：

    2026年6月17日

查看全文
  • Saniya Zarash Zameer Kalandar,
  • Sumaiya Begum Thuraputheen,
  • K. Indra Gandhi,
  • Karthikeyan Ramalingam

摘要

Campylobacter jejuni is a leading cause of bacterial gastroenteritis worldwide, and its growing resistance to antibiotics highlights the urgent need for novel therapeutic agents. This study presents an AI-driven in silico pipeline to identify marine derived bioactive compounds with potential antibacterial activity against C. jejuni. A total of 500 marine natural products were screened, and 58 compounds were selected after filtering using Lipinski’s Rule of Five Veber’s criteria, and ADMET profiling. Fascaplysine, Auranthine, and Ascididemin emerged as the top lead candidates based on favorable pharmacokinetic and safety profiles. Network pharmacology identified six key hub genes TLR4, TNF, MMP3, CYP27B1, ACE, and STAT1 associated with host immune modulation and C. jejuni pathogenesis. These targets are enriched in immune signaling pathways including Toll-like receptor signaling, Necroptosis, and Inflammatory Bowel Disease. Molecular docking using AutoDock 4.2.6 revealed strong binding affinities, with the Fascaplysine-STAT1 complex exhibiting the highest affinity (-9.96 kcal/mol; Ki = 49.53 nM). All three marine lead compounds demonstrated superior binding affinities compared to ciprofloxacin, the clinically approved first-line antibiotic, at all targets for which they were the top-ranked hits; the difference was most pronounced at STAT1, where Fascaplysine (-9.96 kcal/mol) substantially outperformed ciprofloxacin (-6.69 kcal/mol), a difference of 3.27 kcal/mol. A Random Forest regression model trained on dockingmderived features and molecular descriptors achieved an R² of approximately 0.90 outperforming XGBoost, AdaBoost and KNN in binding affinity prediction. A comparative analysis between AutoDock-derived and Random Forest-predicted binding affinities across 30 protein ligand pairs demonstrated strong concordance (mean |ΔAffinity| = 0.66 kcal/mol), confirming the predictive accuracy of the model. The model was integrated into a publicly accessible web application enabling real-time binding affinity prediction using PubChem compound descriptors. Off-target screening via SwissTargetPrediction indicated low predicted cross-reactivity with non-target human proteins for all three lead compounds. Together these findings highlight marine-derived alkaloids as promising antibacterial candidates and provide an AI-enhanced computational framework to accelerate early-stage drug discovery against antibiotic-resistant C. jejuni.