Scholarlink

<p>Chemoresistance in oral squamous cell carcinoma (OSCC) is driven by genetic mutations, altered drug metabolism, and cancer stem cells (CSCs), leading to treatment failure and recurrence. This study aimed to identify and predict key genes associated with Docetaxel resistance using an integrated computational and experimental approach. Gene expression profiles from the GSE175726 dataset were analyzed using GEOquery and WGCNA to construct co-expression networks. Hub genes were identified using CytoHubba and functionally annotated through Enrichr. An artificial neural network (ANN) model was developed, achieving 86% prediction accuracy, an AUC of 0.865, class accuracy of 76.5%, and an F1 score of 66.3%. The model demonstrated high sensitivity but moderate specificity (23.5%), reflecting challenges in non-hub classification. Key hub genes—DUSP2, FXYD3, CYP1B1, MCOLN2, and ITGB2—were validated by qPCR in Docetaxel-resistant OSCC cell lines, confirming their role in CSC regulation and chemoresistance. These findings provide potential therapeutic targets and demonstrate the utility of ANN-assisted discovery in advancing precision oncology for OSCC.</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

Predicting and identifying key genes driving chemoresistance and cancer stemness in oral squamous cell carcinoma

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年6月10日

查看全文
  • Swarnambiga Ayyachamy,
  • Pradeep K. Yadalam,
  • Raghu Nandha Kumar,
  • Charumathi P.,
  • Bharathi Devi M.,
  • Carlos M. Ardila

摘要

Chemoresistance in oral squamous cell carcinoma (OSCC) is driven by genetic mutations, altered drug metabolism, and cancer stem cells (CSCs), leading to treatment failure and recurrence. This study aimed to identify and predict key genes associated with Docetaxel resistance using an integrated computational and experimental approach. Gene expression profiles from the GSE175726 dataset were analyzed using GEOquery and WGCNA to construct co-expression networks. Hub genes were identified using CytoHubba and functionally annotated through Enrichr. An artificial neural network (ANN) model was developed, achieving 86% prediction accuracy, an AUC of 0.865, class accuracy of 76.5%, and an F1 score of 66.3%. The model demonstrated high sensitivity but moderate specificity (23.5%), reflecting challenges in non-hub classification. Key hub genes—DUSP2, FXYD3, CYP1B1, MCOLN2, and ITGB2—were validated by qPCR in Docetaxel-resistant OSCC cell lines, confirming their role in CSC regulation and chemoresistance. These findings provide potential therapeutic targets and demonstrate the utility of ANN-assisted discovery in advancing precision oncology for OSCC.