文献类型：期刊文献

英文题名：Identification of key biomarkers for predicting CAD progression in inflammatory bowel disease via machine-learning and bioinformatics strategies

作者：Tang, Xiaoqi[1];Zhou, Yufei[2,3];Chen, Zhuolin[4];Liu, Chunjiang[5];Wu, Zhifeng[1];Zhou, Yue[5];Zhang, Fan[1];Lu, Xuanyuan[4];Tang, Liming[5,6]

机构：[1]Shaoxing Univ, Sch Med, Shaoxing, Zhejiang, Peoples R China;[2]Fudan Univ, Zhongshan Hosp, Shanghai Inst Cardiovasc Dis, Dept Cardiol, Shanghai, Peoples R China;[3]Fudan Univ, Inst Biomed Sci, Shanghai, Peoples R China;[4]Zhejiang Univ, Shaoxing Peoples Hosp, Dept Orthoped, Sch Med, Shaoxing, Peoples R China;[5]Shaoxing Peoples Hosp, Dept Gen Surg, Div Vasc Surg, Shaoxing, Peoples R China;[6]Shaoxing Univ, Shaoxing Peoples Hosp, Dept Vasc & Hernia Surg, Affiliated Hosp 1, Shaoxing 312000, Peoples R China

年份：2024

卷号：28

期号：6

外文期刊名：JOURNAL OF CELLULAR AND MOLECULAR MEDICINE

收录：SCI-EXPANDED(收录号：WOS:001180833600001)、、Scopus(收录号：2-s2.0-85187159658)、WOS

基金：We sincerely thank the individuals and/or institutions who contributed to this study. The authors would like to thank Sangerbox for providing us an online platform for data processing and analyzes, and thank Dr. Chen and Dr. Yi for their advice in the R operation.

语种：英文

外文关键词：acute myocardial infarction; atherosclerosis progression; bioinformatics; coronary artery disease; inflammatory bowel disease; inflammation and immunity; machine-learning

外文摘要：The study aimed to identify the biomarkers for predicting coronary atherosclerotic lesions progression in patients with inflammatory bowel disease (IBD). Related transcriptome datasets were seized from Gene Expression Omnibus database. IBD-related modules were identified via Weighted Gene Co-expression Network Analysis. The 'Limma' was applied to screen differentially expressed genes between stable coronary artery disease (CAD) and acute myocardial infarction (AMI). Subsequently, we employed protein-protein interaction (PPI) network and three machine-learning strategies to further screen for candidate hub genes. Application of the receiver operating characteristics curve to quantitatively evaluate candidates to determine key diagnostic biomarkers, followed by a nomogram construction. Ultimately, we performed immune landscape analysis, single-gene GSEA and prediction of target-drugs. 3227 IBD-related module genes and 570 DEGs accounting for AMI were recognized. Intersection yielded 85 shared genes and mostly enriched in immune and inflammatory pathways. After filtering through PPI network and multi-machine learning algorithms, five candidate genes generated. Upon validation, CTSD, CEBPD, CYP27A1 were identified as key diagnostic biomarkers with a superior sensitivity and specificity (AUC > 0.8). Furthermore, all three genes were negatively correlated with CD4(+) T cells and positively correlated with neutrophils. Single-gene GSEA highlighted the importance of pathogen invasion, metabolism, immune and inflammation responses during the pathogenesis of AMI. Ten target-drugs were predicted. The discovery of three peripheral blood biomarkers capable of predicting the risk of CAD proceeding into AMI in IBD patients. These identified biomarkers were negatively correlated with CD4(+) T cells and positively correlated with neutrophils, indicating a latent therapeutic target.