A Novel Diagnostic and Subtype Classification Model Based on RNA N6-Methyladenosine Regulators for Behçet's Uveitis.

Abstract

Purpose: To investigate the role of RNA N6-methyladenosine (m⁶A) regulators in the diagnosis and subtype classification of Behçet's uveitis, aiming to establish a novel predictive model and explore distinct molecular patterns for personalized therapeutic approaches.

Methods: Data from the Gene Expression Omnibus GSE209567 dataset comprising 22 Behçet's uveitis patients and 15 controls were analyzed for m⁶A regulator expression. Differentially expressed genes were identified using the "limma" R package, followed by random forest (RF) and support vector machine (SVM) model construction to select critical m6A regulators. A nomogram model was developed for prediction, and consensus clustering identified distinct m⁶A and gene-regulating patterns. Immune cell infiltration analysis was conducted using ssGSEA, and m⁶A scores were computed to quantify molecular patterns.

Results: Eleven m⁶A regulators were significantly differentially expressed. The top four candidate m⁶A regulators (FTO, YTHDF2, CBLL1, and METTL14) were identified to predict the risk of Behçet's uveitis. A nomogram was constructed based on the four candidate m⁶A regulators to visualize the association between the expression levels of the candidate with the risk of onset of Behçet's uveitis. Two distinct m⁶A patterns and gene patterns were identified, validated by consensus clustering. High m⁶A scores were associated with more severe disease stages, with differential immune cell infiltration observed between subtypes. Immune-related genes, such as LRRN3 and DAAM2, were identified as key in differentiating m⁶A patterns.

Conclusion: M⁶A modification plays an important role in the occurrence of uveitis. Distinct m⁶A patterns and gene clusters highlight their potential for early diagnosis and personalized treatment.