Identifying Basement Membrane-Related Diagnostic Biomarkers for Rheumatoid Arthritis Through Machine Learning.

Abstract

Introduction: Rheumatoid Arthritis (RA) is a common autoimmune disease. The Basement Membrane (BM) plays a critical structural role in tissues such as the kidneys and joints, and is often implicated in immune-related diseases. This study aims to explore the complex interactions between RA and the BM, and to identify potential diagnostic biomarkers.

Methods: The data used in this study were sourced from the Gene Expression Omnibus (GEO) database. We screened Differentially Expressed Genes (DEGs) related to BM by comparing RA tissue with normal tissue. Consensus clustering analysis was performed based on the features of the BM in RA tissue samples. Functional enrichment analyses, including Gene Ontology (GO) and KEGG pathway analysis, were performed on genes common to the different clusters. Three machine learning algorithms were used to screen for biomarkers, including LASSO, Random Forest (RF), and Support Vector Machine Recursive Feature Elimination (SVM-RFE). The screening results were validated using the GSE77298 dataset and confirmed via qRT-PCR. In addition, molecular docking technology was applied to predict the binding potential between Schisandrin (SCH) and the candidate biomarkers SEL1L3 and SLAMF8.

Results: By integrating RA transcriptome data with BM-related genes, we identified two BMassociated molecular subtypes of RA, both of which exhibit distinct immune cell infiltration characteristics. By further implementing a combination of LASSO, RF, and SVM-RFE, SEL1L3 and SLAMF8 were identified as candidate diagnostic biomarkers. In the external validation set (GSE77298) and qRT-PCR experiments, both exhibited significant upregulation in RA tissue and excellent diagnostic efficacy. The AUC of SEL1L3 and SLAMF8 was 0.987 and 0.970, respectively, and molecular docking suggests that SCH binds stably to both.

Discussion: The research results indicate that molecular changes related to BM may play an important role in the pathogenesis of RA, particularly through immune regulation and tissue remodeling. Identifying different subtypes of BM-related RA highlights the heterogeneity of RA and may help improve disease classification and personalized diagnosis. SEL1L3 and SLAMF8 can serve as promising biomarkers for early RA detection and provide new insights into BM-related immune mechanisms.

Conclusion: SEL1L3 and SLAMF8 are the first BM-associated RA diagnostic markers identified, providing new targets for early RA diagnosis and investigation of its molecular mechanisms.