BACKGROUND/OBJECTIVES: Rheumatoid arthritis (RA) is a chronic autoimmune disease that causes progressive joint damage and systemic complications. Despite multiple treatment options, many patients fail to achieve sustained remission. Our study aimed to integrate bioinformatics and single-cell RNA-seq analyses to identify potential biomarkers and therapeutic targets and explore bioactive compounds from traditional Chinese medicine (TCM). METHODS: We integrated gene expression quantitative trait loci (eQTL), protein quantitative trait loci (pQTL), and genome-wide association study (GWAS) data for RA using two-sample Mendelian randomization to identify causal druggable genes. Bulk transcriptomics and machine learning were used for candidate gene screening and validation, while single-cell RNA-seq analysis characterized cell type-specific expression and functional relevance. TCM compound screening, molecular docking, and molecular dynamics (MD) simulations were subsequently performed. RESULTS: were identified as RA-associated candidate targets with distinct cell type-specific expression patterns, strong immune associations, and favorable diagnostic performance. Functional analyses linked these genes to immune activation and intercellular communication. In silico analyses prioritized sesamin, (+)-Ganoderic acid Mf, and (24R)-saringosterol as candidate compounds, with the IFNG-(+)-Ganoderic acid Mf complex showing stable behavior during MD simulation. CONCLUSIONS: as candidate biomarkers and druggable targets for RA. Sesamin, (+)-Ganoderic acid Mf, and (24R)-saringosterol warrant further experimental evaluation. These findings provide a basis for future mechanistic and translational studies.