Identification of Diagnostic Biomarkers in Systemic Lupus Erythematosus Based on Bioinformatics Analysis and Machine Learning

• Published in: Frontiers in genetics Vol. 13; p. 865559
• Main Authors: Jiang, Zhihang, Shao, Mengting, Dai, Xinzhu, Pan, Zhixin, Liu, Dongmei
• Format: Journal Article
• Language: English
• Published: Switzerland Frontiers Media S.A 14.04.2022
• Subjects: diagnostic biomarkers; Genetics; immune infiltration; integrated bioinformatics; machine learning; systemic lupus erythematosus; integrated bioinformatics; systemic lupus erythematosus; immune infiltration; machine learning; diagnostic biomarkers
• ISSN: 1664-8021; 1664-8021
• DOI: 10.3389/fgene.2022.865559

Systemic lupus erythematosus (SLE) is a complex autoimmune disease that affects several organs and causes variable clinical symptoms. Exploring new insights on genetic factors may help reveal SLE etiology and improve the survival of SLE patients. The current study is designed to identify key genes involved in SLE and develop potential diagnostic biomarkers for SLE in clinical practice. Expression data of all genes of SLE and control samples in GSE65391 and GSE72509 datasets were downloaded from the Gene Expression Omnibus (GEO) database. A total of 11 accurate differentially expressed genes (DEGs) were identified by the “limma” and “RobustRankAggreg” R package. All these genes were functionally associated with several immune-related biological processes and a single KEGG (Kyoto Encyclopedia of Genes and Genome) pathway of necroptosis. The PPI analysis showed that IFI44, IFI44L, EIF2AK2, IFIT3, IFITM3, ZBP1, TRIM22, PRIC285, XAF1, and PARP9 could interact with each other. In addition, the expression patterns of these DEGs were found to be consistent in GSE39088. Moreover, Receiver operating characteristic (ROC) curves analysis indicated that all these DEGs could serve as potential diagnostic biomarkers according to the area under the ROC curve (AUC) values. Furthermore, we constructed the transcription factor (TF)-diagnostic biomarker-microRNA (miRNA) network composed of 278 nodes and 405 edges, and a drug-diagnostic biomarker network consisting of 218 nodes and 459 edges. To investigate the relationship between diagnostic biomarkers and the immune system, we evaluated the immune infiltration landscape of SLE and control samples from GSE6539. Finally, using a variety of machine learning methods, IFI44 was determined to be the optimal diagnostic biomarker of SLE and then verified by quantitative real-time PCR (qRT-PCR) in an independent cohort. Our findings may benefit the diagnosis of patients with SLE and guide in developing novel targeted therapy in treating SLE patients.