Datasets for identifying biomarkers of rheumatoid arthritis based on bioinformatics, machine learning and experimental validation

Abstract: The pathogenesis of rheumatoid arthritis (RA) is complex, and achieving clinical cure remains challenging. Identifying RA biomarkers is crucial for advancing molecular diagnosis and identifying potential therapeutic targets. Obtain transcriptome data of synovial tissue from RA patients and healthy controls through GEO database and perform differential analysis. Intersection the differentially expressed genes, RA related genes screened by weighted gene co expression network analysis, and neutrophil extracellular trap (NETs) related genes searched in literature database to obtain RA related NETs genes. Obtain immune cells with differences between the experimental group and the control group through ssGSEA analysis, and identify immune cells related to differential genes through immune cell correlation analysis; LASSO regression, support vector machine, and random forest tree were used to screen RA feature genes, and the intersection of the three was taken to obtain RA feature genes. The diagnostic value of the feature genes was evaluated and validated. Perform typing and PCA analysis on the samples, and perform immune cell differential analysis, pathway enrichment analysis, and drug enrichment analysis on the typing results, and construct a gene drug regulatory network. Finally, a rheumatoid arthritis (CIA) rat model was constructed for validation. Through identification, 10 RA related NETs genes were identified, including immunoglobulin heavy chain constant gamma 1, chemokine ligand CXC 13 (CXCL13), multi ligand proteoglycan 1 (SDC1), and so on. Immune infiltration analysis showed that these genes were significantly associated with activated B cells, memory B cells, CD4 T cells, CD8 T cells, and other immune cells. And through three machine learning algorithms, CXCL13 and SDC1 were identified as RA feature genes with diagnostic value. The samples can be divided into two subtypes, C1 and C2. Activated B cells, activated CD4 T cells, activated CD8 T cells and other immune cells have significant differences between the two subtypes. The signaling pathways of primary immunodeficiency, antigen processing and presentation signals, and allograft rejection have significant differences between the two subtypes. Drugs such as methylphenidate and isoguanine are significantly enriched with characteristic genes. Finally, HE staining, RT qPCR, and Western blot showed that compared with normal rats, CIA rats had a higher degree of immune cell infiltration in synovial tissue, and higher expression levels of CXCL13 and SDC1. The above results indicate that CXCL13 and SDC1 are characteristic genes of RA, highly expressed in RA synovial tissue, promoting the pathogenesis of RA and providing new insights for the diagnosis and immunotherapy strategies of RA.