Scholarlink

Objective&#xa0;and design <p>This study aimed to identify robust diagnostic biomarkers and characterize molecular subtypes of chronic rhinosinusitis (CRS) by integrating multi-omics data with machine learning, utilizing a case–control design with patient samples.</p> Material or subjects <p>The analysis incorporated gene expression data from bulk and single-cell RNA sequencing datasets; validation experiments used clinical samples from human CRS patients and control subjects.</p> Treatment <p>Not applicable.</p> Methods <p>We analyzed differentially expressed genes and immune cell infiltration from transcriptomic data, employed machine learning algorithms to select diagnostic genes and build a predictive model, and validated key targets using quantitative real-time PCR and dual-fluorescence immunohistochemical staining.</p> Results <p>Machine learning identified six inflammation-related genes (SPI1, IFITM3, ITGAM, BCL2A1, HLA-DPB1, PLA2G7) as a diagnostic signature, with predictive models demonstrating strong diagnostic performance. Validation confirmed significant differential expression of BCL2A1 and PLA2G7 in CRS patient samples compared to controls.</p> Conclusions <p>This integrative analysis highlights the utility of computational approaches for discovering CRS biomarkers and subtypes, implicating specific genes in inflammation-associated pathways and paving the way for precision diagnostics and targeted therapies.</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

Comprehensive immunological and molecular analysis revealed inflammation-related diagnostic signatures in chronic rhinosinusitis

  • 期刊论文

  • 发表日期：

    2026年3月17日

查看全文
  • Changchun Yu,
  • Zhaonan Xu,
  • Xianyan Zhao,
  • Siyu Gu,
  • Xuan Kan

摘要

Objective and design

This study aimed to identify robust diagnostic biomarkers and characterize molecular subtypes of chronic rhinosinusitis (CRS) by integrating multi-omics data with machine learning, utilizing a case–control design with patient samples.

Material or subjects

The analysis incorporated gene expression data from bulk and single-cell RNA sequencing datasets; validation experiments used clinical samples from human CRS patients and control subjects.

Treatment

Not applicable.

Methods

We analyzed differentially expressed genes and immune cell infiltration from transcriptomic data, employed machine learning algorithms to select diagnostic genes and build a predictive model, and validated key targets using quantitative real-time PCR and dual-fluorescence immunohistochemical staining.

Results

Machine learning identified six inflammation-related genes (SPI1, IFITM3, ITGAM, BCL2A1, HLA-DPB1, PLA2G7) as a diagnostic signature, with predictive models demonstrating strong diagnostic performance. Validation confirmed significant differential expression of BCL2A1 and PLA2G7 in CRS patient samples compared to controls.

Conclusions

This integrative analysis highlights the utility of computational approaches for discovering CRS biomarkers and subtypes, implicating specific genes in inflammation-associated pathways and paving the way for precision diagnostics and targeted therapies.