Background
RNA methylation modifications, such as N1-methyladenosine/N6-methyladenosine /N5-methylcytosine (m1A/m6A/m5C), are the most common RNA modifications and are crucial for a number of biological processes. Nonetheless, the role of RNA methylation modifications of m1A/m6A/m5C in the pathogenesis of renal interstitial fibrosis (RIF) remains incompletely understood.
Methods
Firstly, we downloaded 2 expression datasets from the GEO database, namely GSE22459 and GSE76882. In a differential analysis of these datasets between patients with and without RIF, we selected 33 methylation-related genes (MRGs). We then applied a PPI network, LASSO analysis, SVM-RFE algorithm, and RF algorithm to identify key MRGs.
Results
We eventually obtained five candidate MRGs (WTAP, ALKBH5, YTHDF2, RBMX, and ELAVL1) to forecast the risk of RIF. We created a nomogram model derived from five key MRGs, which revealed that the nomogram model may be advantageous to patients. Based on the selected five significant MRGs, patients with RIF were classified into two MRG patterns using consensus clustering, and the correlation between the five MRGs, the two MRG patterns, and the genetic pattern with immune cell infiltration was shown. Moreover, we conducted GO and KEGG analyses on 768 DEGs between MRG clusters A and B to look into their different involvement in RIF. To measure the MRG patterns, a PCA algorithm was developed to determine MRG scores for each sample. The MRG scores of the patients in cluster B were higher than those in cluster A.
Conclusions
Ultimately, we concluded that cluster A in the two MRG patterns identified on these five key m1A/m6A/m5C regulators may be associated with RIF.