Scholarlink

Background <p>Mitophagy has been implicated in the pathogenesis of rheumatoid arthritis (RA), particularly in fibroblast-like synoviocytes (FLSs), yet its regulatory landscape and diagnostic relevance remain incompletely understood. This study aimed to identify mitophagy-associated genes (MRGs) in RA and to explore their potential diagnostic value using integrative bioinformatics and interpretable machine learning approaches.</p> Methods <p>Transcriptomic datasets of RA patients and healthy controls were retrieved from the Gene Expression Omnibus (GEO). Mitophagy-related differentially expressed genes were identified using limma-based differential analysis and functional enrichment. Multiple supervised machine learning algorithms were applied within a unified modeling framework to prioritize feature genes and evaluate diagnostic performance. NRAS expression was further examined in TNF-α–stimulated MH7A cells and a collagen-induced arthritis (CIA) rat model. In addition, exploratory drug–gene interaction analysis and molecular docking were performed to identify potential compounds targeting NRAS-related pathways.</p> Results <p>Eight mitophagy-related differentially expressed genes were identified, among which NRAS and SREBF1 were consistently selected as key features across multiple machine learning models, showing robust discriminative performance (AUC &gt; 0.90). A logistic regression model based on these genes demonstrated stable diagnostic performance in internal and external validation cohorts. NRAS was significantly upregulated in RA and was experimentally confirmed in vitro and in vivo. Unsupervised clustering further revealed distinct mitophagy-associated RA subtypes with differential immune infiltration patterns and functional pathway enrichment. Drug–gene interaction analysis and molecular docking suggested Binimetinib and Sotorasib as candidate compounds with potential relevance to NRAS signaling.</p> Conclusion <p>NRAS and SREBF1 were identified as mitophagy-associated genes with diagnostic relevance in RA. NRAS, in particular, showed consistent expression changes and experimental validation, supporting its potential involvement in mitophagy-related regulatory networks. These findings provide integrative and hypothesis-generating insights into mitophagy-associated mechanisms in RA and may inform future diagnostic and therapeutic studies.</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

NRAS and SREBF1 identified as mitophagy-associated risk genes for rheumatoid arthritis through interpretable machine learning and experimental validation studies

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年2月7日

查看全文
  • Hongman Li,
  • Peng Liu,
  • Xiao Han,
  • Shengyong Si,
  • ZhiJing Li,
  • Chun Zhao,
  • Xue Zhao,
  • Xue Zhou,
  • Mengdie Liu

摘要

Background

Mitophagy has been implicated in the pathogenesis of rheumatoid arthritis (RA), particularly in fibroblast-like synoviocytes (FLSs), yet its regulatory landscape and diagnostic relevance remain incompletely understood. This study aimed to identify mitophagy-associated genes (MRGs) in RA and to explore their potential diagnostic value using integrative bioinformatics and interpretable machine learning approaches.

Methods

Transcriptomic datasets of RA patients and healthy controls were retrieved from the Gene Expression Omnibus (GEO). Mitophagy-related differentially expressed genes were identified using limma-based differential analysis and functional enrichment. Multiple supervised machine learning algorithms were applied within a unified modeling framework to prioritize feature genes and evaluate diagnostic performance. NRAS expression was further examined in TNF-α–stimulated MH7A cells and a collagen-induced arthritis (CIA) rat model. In addition, exploratory drug–gene interaction analysis and molecular docking were performed to identify potential compounds targeting NRAS-related pathways.

Results

Eight mitophagy-related differentially expressed genes were identified, among which NRAS and SREBF1 were consistently selected as key features across multiple machine learning models, showing robust discriminative performance (AUC > 0.90). A logistic regression model based on these genes demonstrated stable diagnostic performance in internal and external validation cohorts. NRAS was significantly upregulated in RA and was experimentally confirmed in vitro and in vivo. Unsupervised clustering further revealed distinct mitophagy-associated RA subtypes with differential immune infiltration patterns and functional pathway enrichment. Drug–gene interaction analysis and molecular docking suggested Binimetinib and Sotorasib as candidate compounds with potential relevance to NRAS signaling.

Conclusion

NRAS and SREBF1 were identified as mitophagy-associated genes with diagnostic relevance in RA. NRAS, in particular, showed consistent expression changes and experimental validation, supporting its potential involvement in mitophagy-related regulatory networks. These findings provide integrative and hypothesis-generating insights into mitophagy-associated mechanisms in RA and may inform future diagnostic and therapeutic studies.