Identification of a ferroptosis-related gene pair biomarker with immune infiltration landscapes in ischemic stroke: a bioinformatics-based comprehensive study

• Published in: BMC genomics 2022-01, Vol.23 (1), p.59-59, Article 59
• Main Authors: Fan, Jiaxin, Chen, Mengying, Cao, Shuai, Yao, Qingling, Zhang, Xiaodong, Du, Shuang, Qu, Huiyang, Cheng, Yuxuan, Ma, Shuyin, Zhang, Meijuan, Huang, Yizhou, Zhang, Nan, Shi, Kaili, Zhan, Shuqin
• Format: Article
• Language: English
• Subjects: Algorithms; Alzheimer's disease; Apoptosis; Bioinformatics; Biomarker; Biomarkers; Brain Ischemia - genetics; Care and treatment; Cell death; Computational Biology; Correlation analysis; Data mining; Datasets; Diagnosis; Diagnostic systems; Ferroptosis; Gene expression; Gene pair; Genes; Genetic Markers; Genomics; Health aspects; Hemorrhage; Humans; Immune infiltration; Infiltration; Ischemia; Ischemic Stroke; Learning algorithms; Lipid peroxidation; Lipids; Machine learning; Mast cells; Metadata; Methods; Microenvironments; MicroRNAs; miRNA; Morphology; mRNA; Online data bases; Peroxidation; Plasma cells; Ribonucleic acid; Risk factors; RNA; Stroke; Stroke (Disease); Stroke - genetics; Support vector machines; Therapeutic targets
• ISSN: 1471-2164; 1471-2164
• DOI: 10.1186/s12864-022-08295-0

Ischemic stroke (IS) is a principal contributor to long-term disability in adults. A new cell death mediated by iron is ferroptosis, characterized by lethal aggregation of lipid peroxidation. However, a paucity of ferroptosis-related biomarkers early identify IS until now. This study investigated potential ferroptosis-related gene pair biomarkers in IS and explored their roles in immune infiltration. In total, we identified 6 differentially expressed ferroptosis-related genes (DEFRGs) in the metadata cohort. Of these genes, 4 DEFRGs were incorporated into the competitive endogenous RNA (ceRNA) network, including 78 lncRNA-miRNA and 16 miRNA-mRNA interactions. Based on relative expression values of DEFRGs, we constructed gene pairs. An integrated scheme consisting of machine learning algorithms, ceRNA network, and gene pair was proposed to screen the key DEFRG biomarkers. The receiver operating characteristic (ROC) curve witnessed that the diagnostic performance of DEFRG pair CDKN1A/JUN was superior to that of single gene. Moreover, the CIBERSORT algorithm exhibited immune infiltration landscapes: plasma cells, resting NK cells, and resting mast cells infiltrated less in IS samples than controls. Spearman correlation analysis confirmed a significant correlation between plasma cells and CDKN1A/JUN (CDKN1A: r = - 0.503, P