Enhancing network analysis with supervised machine learning and mendelian randomization with unsupervised machine learning to identify core phase separation biomarkers in autoimmune insulin receptoropathy

IF 0.7 Q4 GENETICS & HEREDITY

Human Gene Pub Date : 2025-07-23 DOI:10.1016/j.humgen.2025.201455

Chuyu Liang , Zhaoxia Yu , Qiuyi Liang , Ziran Zeng , Rongguan Ma , Wenyan Xie , Xiao Zhu

{"title":"Enhancing network analysis with supervised machine learning and mendelian randomization with unsupervised machine learning to identify core phase separation biomarkers in autoimmune insulin receptoropathy","authors":"Chuyu Liang , Zhaoxia Yu , Qiuyi Liang , Ziran Zeng , Rongguan Ma , Wenyan Xie , Xiao Zhu","doi":"10.1016/j.humgen.2025.201455","DOIUrl":null,"url":null,"abstract":"<div><h3>Background</h3><div>Recent research has focused on the link between phase separation genes and immunity, alongside their potential role in insulin signaling modulation. Autoimmune insulin receptoropathy (AIR), characterized by sporadic hypoglycemia, lacks reliable molecular markers for early detection.</div></div><div><h3>Methods</h3><div>Phase separation genes associated with AIR were analyzed using differential analysis and Weighted Gene Co-expression Network Analysis (WGCNA). Gene Set Enrichment Analysis (GSEA), Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Disease Ontology (DO) explored biological differences. A protein-protein interaction (PPI) network and machine learning (SVM-REF, RandomForest) identified core phase separation genes. Functional insights were gained through correlation, differential expression, and single-gene GSEA analyses. Marker gene activity was assessed via single-sample GSEA. Mendelian randomization (MR) examined potential causal links between DO results and the disease, validating associations with phase separation genes. Unsupervised machine learning reinforced the findings.</div></div><div><h3>Results</h3><div>Differential gene expression and WGCNA identified 2944 differentially expressed genes and 18 co-expression modules in AIR. The darkturquoise module, showing a potential inverse relationship with disease status, was selected for further analysis. GSEA revealed up-regulated pathways such as Inositol phosphate metabolism and down-regulated pathways like Drug metabolism. PPI network and machine learning analyses identified 10 core genes closely linked to AIR, demonstrating significant predictive capability and potential as early diagnostic biomarkers.</div></div><div><h3>Conclusions</h3><div>The phase separation genes linked to AIR show strong disease associations, offering potential for early prediction and improved clinical management.</div></div>","PeriodicalId":29686,"journal":{"name":"Human Gene","volume":"45 ","pages":"Article 201455"},"PeriodicalIF":0.7000,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Human Gene","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2773044125000816","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"GENETICS & HEREDITY","Score":null,"Total":0}

引用次数: 0

Abstract

Background

Recent research has focused on the link between phase separation genes and immunity, alongside their potential role in insulin signaling modulation. Autoimmune insulin receptoropathy (AIR), characterized by sporadic hypoglycemia, lacks reliable molecular markers for early detection.

Methods

Phase separation genes associated with AIR were analyzed using differential analysis and Weighted Gene Co-expression Network Analysis (WGCNA). Gene Set Enrichment Analysis (GSEA), Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and Disease Ontology (DO) explored biological differences. A protein-protein interaction (PPI) network and machine learning (SVM-REF, RandomForest) identified core phase separation genes. Functional insights were gained through correlation, differential expression, and single-gene GSEA analyses. Marker gene activity was assessed via single-sample GSEA. Mendelian randomization (MR) examined potential causal links between DO results and the disease, validating associations with phase separation genes. Unsupervised machine learning reinforced the findings.

Results

Differential gene expression and WGCNA identified 2944 differentially expressed genes and 18 co-expression modules in AIR. The darkturquoise module, showing a potential inverse relationship with disease status, was selected for further analysis. GSEA revealed up-regulated pathways such as Inositol phosphate metabolism and down-regulated pathways like Drug metabolism. PPI network and machine learning analyses identified 10 core genes closely linked to AIR, demonstrating significant predictive capability and potential as early diagnostic biomarkers.

Conclusions

The phase separation genes linked to AIR show strong disease associations, offering potential for early prediction and improved clinical management.

Abstract Image

查看原文本刊更多论文

加强网络分析与监督机器学习和孟德尔随机化与无监督机器学习，以确定自身免疫性胰岛素受体病变的核心相分离生物标志物

最近的研究主要集中在相分离基因与免疫之间的联系，以及它们在胰岛素信号调节中的潜在作用。自身免疫性胰岛素受体病变（AIR），以散发性低血糖为特征，缺乏可靠的早期检测分子标记。方法采用差异分析和加权基因共表达网络分析（WGCNA）对AIR相关相分离基因进行分析。基因集富集分析（GSEA）、基因本体（GO）、京都基因与基因组百科全书（KEGG）和疾病本体（DO）探讨了生物学差异。蛋白质-蛋白质相互作用（PPI）网络和机器学习（SVM-REF, RandomForest）确定了核心相分离基因。通过相关性、差异表达和单基因GSEA分析获得功能见解。通过单样本GSEA评估标记基因活性。孟德尔随机化（MR）研究了DO结果与疾病之间的潜在因果关系，验证了与相分离基因的关联。无监督机器学习强化了这一发现。结果差异基因表达和WGCNA鉴定出2944个差异基因和18个共表达模块。暗绿松石模块显示了与疾病状态的潜在反比关系，因此被选中进行进一步分析。GSEA显示肌醇磷酸代谢等上调通路和药物代谢等下调通路。PPI网络和机器学习分析确定了10个与AIR密切相关的核心基因，显示出显著的预测能力和作为早期诊断生物标志物的潜力。结论与AIR相关的相分离基因表现出较强的疾病相关性，为早期预测和改善临床管理提供了潜力。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊