Machine Learning Identification of Metabolism-Related Biomarkers with Diagnostic Potential for Gastric Cancer: Multi-Dimensional Transcriptomic Validation

利用机器学习识别具有胃癌诊断潜力的代谢相关生物标志物:多维转录组学验证

阅读:1

Abstract

BACKGROUND: An increasing body of evidence suggests an association between metabolic syndrome and gastric cancer. However, the shared genetic signatures and underlying molecular mechanisms between them remain to be elucidated. METHODS: We obtained transcriptomic data for gastric cancer and metabolic syndrome from the GEO, TCGA, and GTEx databases. Using the Limma and WGCNA algorithms respectively, we identified differential genes and co-expression module genes related to metabolic syndrome and gastric cancer. Lasso and SVM were employed to further screen for hub genes, while XGBoost was utilized to enhance the diagnostic value of these hub genes. CIBERSORT and GSVA were applied to assess the correlation among hub genes for immune infiltration and metabolic scores. Single-cell and spatial transcriptomic analyses were conducted to explore cell subpopulations and tissue distribution of hub genes in gastric cancer. We used qPCR experiments to detect expression differences of hub genes between gastric cancer tissues and normal tissues. RESULTS: CSE1L, IL32, and CCDC86 were identified as shared hub genes between metabolic syndrome and gastric cancer. These genes were significantly associated with immune cell infiltration and dysregulated metabolic pathways. Single-cell analysis revealed elevated glycolysis across gastric cancer cell subpopulations, accompanied by enhanced cell-cell interactions. Spatial transcriptomic analysis confirmed the upregulation of hub genes in tumor regions. qPCR further verified significantly higher mRNA expression levels of these genes in gastric cancer tissues than in adjacent normal tissues. CONCLUSION: CSE1L, IL32, and CCDC86 may represent potential metabolism-related biomarkers associated with gastric cancer and metabolic syndrome. These findings provide additional insight into the molecular links between the two conditions and may support future mechanistic studies and larger-scale clinical validation.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。