Scholarlink

<p>This study aimed to develop a predictive model for mucus-binding proteins using machine learning and to experimentally evaluate the anti-cariogenic effects of selected probiotic strains. In silico, a computational method was established utilizing Support Vector Machine (SVM) and AdaBoost algorithms with pseudo amino acid composition (PseAAC) for protein sequence representation. The predictive model achieved high accuracy. Specifically, the SVM model demonstrated 94% accuracy, 96% sensitivity, 91% specificity, and an 88% Matthews correlation coefficient (MCC) on a labeled test dataset. In vitro experiments assessed the antimicrobial activity and anti-biofilm formation effects of various probiotic strains against <i>Streptococcus mutans</i>. <i>Lactobacillus plantarum 1058</i> exhibited the highest inhibitory effect on <i>S. mutans</i> growth, reducing the bacterial count to 4.3 log CFU/ml after 24&#xa0;h, while <i>Bifidobacterium adolescentis 1536</i> inhibited it the least (5.4 log CFU/ml). Furthermore, <i>L. plantarum 1058</i> demonstrated the highest inhibition of <i>S. mutans</i> biofilm formation (98.68%), whereas <i>Bifidobacterium animalis subsp. lactis</i> showed the lowest inhibition (75.18%). These findings suggest that the developed computational model effectively predicts mucus-binding proteins and the evaluated probiotic strains hold promise for inhibiting <i>S. mutans</i> growth and biofilm formation, thus offering promising strategies for maintaining oral health and preventing dental caries.</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

In silico and in vitro evaluation of mucus-binding proteins from probiotics against Streptococcus mutans

  • 期刊论文

  • 发表日期：

    2025年8月14日

查看全文
  • Ghazaleh Sheikhi,
  • Soheil Shajari,
  • Sepehr Nouri,
  • Hassan Mohabatkar,
  • Mandana Behbahani

摘要

This study aimed to develop a predictive model for mucus-binding proteins using machine learning and to experimentally evaluate the anti-cariogenic effects of selected probiotic strains. In silico, a computational method was established utilizing Support Vector Machine (SVM) and AdaBoost algorithms with pseudo amino acid composition (PseAAC) for protein sequence representation. The predictive model achieved high accuracy. Specifically, the SVM model demonstrated 94% accuracy, 96% sensitivity, 91% specificity, and an 88% Matthews correlation coefficient (MCC) on a labeled test dataset. In vitro experiments assessed the antimicrobial activity and anti-biofilm formation effects of various probiotic strains against Streptococcus mutans. Lactobacillus plantarum 1058 exhibited the highest inhibitory effect on S. mutans growth, reducing the bacterial count to 4.3 log CFU/ml after 24 h, while Bifidobacterium adolescentis 1536 inhibited it the least (5.4 log CFU/ml). Furthermore, L. plantarum 1058 demonstrated the highest inhibition of S. mutans biofilm formation (98.68%), whereas Bifidobacterium animalis subsp. lactis showed the lowest inhibition (75.18%). These findings suggest that the developed computational model effectively predicts mucus-binding proteins and the evaluated probiotic strains hold promise for inhibiting S. mutans growth and biofilm formation, thus offering promising strategies for maintaining oral health and preventing dental caries.