Combination of Mendelian Randomization and Transcriptomes to Identify Causal Genes Associated With IgAN

Background IgA nephropathy (IgAN) represents the most common form of primary glomerulonephritis worldwide, contributing substantially to chronic renal impairment and end‐stage kidney failure. The present study utilized bioinformatics methodologies to screen for pivotal genes involved in IgAN development. Methods Gene expression datasets GSE93798 and GSE104948 were retrieved from the Gene Expression Omnibus (GEO) repository. Differentially expressed genes (DEGs) between IgAN patients and healthy controls were identified using the GSE93798 dataset. Mendelian randomization analysis was subsequently applied to determine genes demonstrating causal associations with IgAN among the identified DEGs. Feature gene selection was achieved by extracting overlapping genes identified through both the least absolute shrinkage and selection operator (LASSO) regression modeling and random forest machine learning approaches. The receiver operating characteristic (ROC) curve methodology was employed to evaluate diagnostic capabilities, while a nomogram‐based prediction tool was developed. To investigate the regulatory pathways and functional properties of selected genes, a transcription factor (TF)–microRNA (miRNA)‐gene interaction network was constructed alongside gene set enrichment analysis (GSEA). Reverse transcription quantitative polymerase chain reaction (RT‐qPCR) validation of signature gene expression was conducted using clinical tissue samples. Results Twelve IgAN–DEGs were significantly associated with IgAN, and five characteristic genes ( MGST1 , FOSB , TSC22D3 , CSF1R , and CRIP1 ) were identified. The nomogram demonstrated robust diagnostic precision for IgAN, as evidenced by calibration and decision curve assessments. GSEA showed that the characteristic genes were enriched in amino acid metabolism–related pathways. In the regulatory network, the associations of hsa‐miR‐760 with TSC22D3 , FOSB , MGST1 , and CRIP1 , associations of hsa‐miR‐512‐3p with FOSB , MGST1 , and CSF1R , and the connections of CEBPB and MYC with eight miRNAs each suggested their importance. Notably, three genes, including TSC22D3, CSF1R , and CRIP1 , exhibited marked expression differences when comparing IgAN patients with healthy controls. Conclusion Through systematic screening, five IgAN‐related signature genes were identified, based on which a diagnostic nomogram model was subsequently established. These findings could facilitate future studies on the pathological mechanisms of IgAN and improve its diagnostic accuracy.