WGCNA and machine learning analysis identifi ed SAMD9 and IFIT3 as primary Sjögren's Syndrome key genes

Abstract

Background: Current treatments for primary Sjögren's Syndrome (pSS) are with limited effect, partially due to the heterogeneity and uncleared mechanism. Methods: We got GSE40568 (Japan) and GSE40611 (USA), and analyzed them with WGCNA to find key Differentially expressed genes (DEGs) between pSS and healthy salivary glands (SG). Key pSS genes (KPGs) were further selected through 3 machine-learning methods. The expression of KPGs was validated via two other GEO datasets (GSE127952 and GSE154926). Infiltrated immune cells, ceRNA network, and potential compounds were explored. Results: Our study identified 376 DEGs from the pSS patients, with 186 genes located in the “plum2'' module, showing the strongest correlation with clinical characteristics. SAMD9 and IFIT3 emerged as KPGs with excellent diagnostic potential. SAMD9 demonstrated close association with immune cell infiltration. We constructed a lncRNA-miRNA-mRNA network comprising 2 KPGs, 12 miRNAs, 124 lncRNAs, and potential therapeutic targets. Conclusion: In the investigation of pSS public datasets, our study revealed two potential critical mediators in the pathological process of pSS salivary glands, namely SAMD9 and IFIT3. Furthermore, we put forth a hypothesis regarding the ceRNA network and made predictions regarding potential therapeutic drugs targeting these two genes.

Keywords

• Primary Sjögren's syndrome
• SAMD9
• IFIT3
• Machine-learning
• WGCNA