Identification of Key CircRNAs Related to Pulmonary Tuberculosis Based on Bioinformatics Analysis

• Published in: BioMed research international Vol. 2022; pp. 1717784 - 15
• Main Authors: Yuan, Qin, Wen, Zilu, Yang, Ke, Zhang, Shulin, Zhang, Ning, Song, Yanzheng, Chen, Fuxue
• Format: Journal Article
• Language: English
• Published: United States Hindawi 04.04.2022; John Wiley & Sons, Inc; Hindawi Limited
• Subjects: Bioinformatics; Biomarkers; Cdc4 protein; Chromosome 5; Computational Biology; Diagnosis; Encyclopedias; Factor analysis; Feature selection; Gene expression; Gene Expression Profiling; Gene Regulatory Networks; Genes; Genomes; Hepatitis; Humans; Identification and classification; Infectious diseases; Learning algorithms; Lung cancer; Machine learning; Medical diagnosis; Metabolism; MicroRNAs - genetics; MicroRNAs - metabolism; miRNA; Patients; Protein interaction; Proteins; Pulmonary tuberculosis; Regulatory mechanisms (biology); RNA; RNA, Circular - genetics; RNA, Messenger - genetics; RNA, Messenger - metabolism; Tuberculosis; Tuberculosis, Pulmonary - genetics; China
• DOI: 10.1155/2022/1717784
• ISSN: 2314-6133

Pulmonary tuberculosis (TB) is a chronic infectious disease that is caused by respiratory infections, principally Mycobacterium tuberculosis. Increasingly, studies have shown that circular (circ)RNAs play regulatory roles in different diseases through different mechanisms. However, their roles and potential regulatory mechanisms in pulmonary TB remain unclear. In this study, we analyzed circRNA sequencing data from adjacent normal and diseased tissues from pulmonary TB patients and analyzed the differentially expressed genes. We then constructed machine learning models and used single-factor analysis to identify hub circRNAs. We downloaded the pulmonary TB micro (mi)RNA (GSE29190) and mRNA (GSE83456) gene expression datasets from the Gene Expression Omnibus database and performed differential expression analysis to determine the differentially expressed miRNAs and mRNAs. We also constructed a circRNA–miRNA–mRNA interaction network using Cytoscape. Gene Ontology enrichment analysis and Kyoto Encyclopedia of Genes and Genomes pathway analysis were used to predict the biological functions of the identified RNAs and determine hub genes. Then, the STRING database and cytoHubba were used to construct protein-protein interaction networks. The results showed 125 differentially expressed circRNAs in the adjacent normal and diseased tissues of pulmonary TB patients. Among them, we identified three hub genes associated with the development of pulmonary TB: hsa_circ_0007919 (upregulated), hsa_circ_0002419 (downregulated), and hsa_circ_0005521 (downregulated). Through further screening, we determined 16 mRNAs of potential downstream genes for hsa-miR-409-5p and hsa_circ_0005521 and established an interaction network. This network may have important roles in the occurrence and development of pulmonary TB. We constructed a model with 100% prediction accuracy by machine learning and single-factor analysis. We constructed a protein-protein interaction network among the top 50 hub mRNAs, with FBXW7 scoring the highest and SOCS3 the second highest. These results may provide a new reference for the identification of candidate markers for the early diagnosis and treatment of pulmonary TB.