Scholarlink

Background <p>Alzheimer’s disease (AD) is a progressive neurodegenerative disorder with complex underlying mechanisms. PANoptosis, a newly defined form of programmed cell death that integrates pyroptosis, apoptosis, and necroptosis, may play a crucial role in AD pathogenesis. However, the involvement of PANoptosis-related genes in AD remains unclear.</p> Methods <p>We analyzed single-cell RNA-seq data (GSE181279) to identify differentially expressed genes (scDEGs) between AD patients and normal controls. PANoptosis-associated genes (PAGs), curated from published studies, were intersected with differentially expressed genes (DEGs) from GSE85426 to identify differentially expressed PAGs (DE-PAGs). Weighted gene co-expression network analysis was performed to identify key gene modules. Candidate genes were identified by overlapping scDEGs, DEGs, and module genes. Hub genes were screened via three machine learning algorithms: least absolute shrinkage and selection operator (LASSO), support vector machine (SVM), and random forest (RF). Genes with consistent expression across GSE85426 and GSE48350 were considered potential biomarkers. These were evaluated by receiver operating characteristic analysis and incorporated into a nomogram. Gene set enrichment analysis was used to explore associated pathways. Immune infiltration analysis was used to assess the biomarkers’ roles in the immune microenvironment and identify potential therapeutic targets. Finally, qRT-PCR was performed to validate biomarker expression in clinical samples.</p> Results <p>Overlapping 987 scDEGs, 991 DEGs, and 5327 module genes yielded 27 candidate genes. LASSO, SVM, and RF analyses identified eight hub genes, among which five (<i>BACH2</i>, <i>CKAP4</i>, <i>DDIT4</i>, <i>GGNBP2</i>, and <i>ZFP36L2</i>) were ultimately validated as biomarkers. A nomogram based on these genes showed good predictive performance (area under the curve (AUC) = 0.779). Seven immune cell types differed significantly between the AD and control groups, with T follicular helper cells strongly correlated with most biomarkers except CKAP4 (cor &gt; 0.36, <i>p</i> &lt; 0.001). Several Aβ- and tau-related genes and immune factors also showed significant associations (|cor|&gt; 0.3, <i>p</i> &lt; 0.05). These biomarkers were further linked to AD and other functional pathways. qRT-PCR was used to validate the transcriptomic findings, with the exception of BACH2.</p> Conclusions <p>This study identified five novel PANoptosis-related biomarkers with diagnostic and therapeutic potential in AD. These findings provide a theoretical basis for future clinical research and may contribute to improved AD diagnosis and treatment strategies.</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

Identification and validation of PANoptosis-related biomarkers in Alzheimer’s disease via single-cell RNA sequencing and machine learning

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年11月25日

查看全文
  • Leishen Li,
  • Fangfang Xu,
  • Hongyan Duan,
  • Junjia Qi,
  • Jingyi Zhang,
  • Kai Ma

摘要

Background

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder with complex underlying mechanisms. PANoptosis, a newly defined form of programmed cell death that integrates pyroptosis, apoptosis, and necroptosis, may play a crucial role in AD pathogenesis. However, the involvement of PANoptosis-related genes in AD remains unclear.

Methods

We analyzed single-cell RNA-seq data (GSE181279) to identify differentially expressed genes (scDEGs) between AD patients and normal controls. PANoptosis-associated genes (PAGs), curated from published studies, were intersected with differentially expressed genes (DEGs) from GSE85426 to identify differentially expressed PAGs (DE-PAGs). Weighted gene co-expression network analysis was performed to identify key gene modules. Candidate genes were identified by overlapping scDEGs, DEGs, and module genes. Hub genes were screened via three machine learning algorithms: least absolute shrinkage and selection operator (LASSO), support vector machine (SVM), and random forest (RF). Genes with consistent expression across GSE85426 and GSE48350 were considered potential biomarkers. These were evaluated by receiver operating characteristic analysis and incorporated into a nomogram. Gene set enrichment analysis was used to explore associated pathways. Immune infiltration analysis was used to assess the biomarkers’ roles in the immune microenvironment and identify potential therapeutic targets. Finally, qRT-PCR was performed to validate biomarker expression in clinical samples.

Results

Overlapping 987 scDEGs, 991 DEGs, and 5327 module genes yielded 27 candidate genes. LASSO, SVM, and RF analyses identified eight hub genes, among which five (BACH2, CKAP4, DDIT4, GGNBP2, and ZFP36L2) were ultimately validated as biomarkers. A nomogram based on these genes showed good predictive performance (area under the curve (AUC) = 0.779). Seven immune cell types differed significantly between the AD and control groups, with T follicular helper cells strongly correlated with most biomarkers except CKAP4 (cor > 0.36, p < 0.001). Several Aβ- and tau-related genes and immune factors also showed significant associations (|cor|> 0.3, p < 0.05). These biomarkers were further linked to AD and other functional pathways. qRT-PCR was used to validate the transcriptomic findings, with the exception of BACH2.

Conclusions

This study identified five novel PANoptosis-related biomarkers with diagnostic and therapeutic potential in AD. These findings provide a theoretical basis for future clinical research and may contribute to improved AD diagnosis and treatment strategies.