Exploration of the shared diagnostic genes and mechanisms between periodontitis and primary Sjögren's syndrome by integrated comprehensive bioinformatics analysis and machine learning

Shaoru Wang,Qimin Wang,Kai Zhao,Shengchao Zhang,Zhenggang Chen
DOI: https://doi.org/10.1016/j.intimp.2024.112899
2024-11-15
Abstract:Background: Accumulating evidence has showed a bidirectional link between periodontitis (PD) and primary Sjögren's syndrome (pSS), but the mechanisms of their occurrence remain unclear. Hence, this study aimed to investigate the shared diagnostic genes and potential mechanisms between PD and pSS using bioinformatics methods. Methods: Gene expression data for PD and pSS were acquired from the Gene Expression Omnibus (GEO) database. Differential expression genes (DEGs) analysis and weighted gene co-expression network analysis (WGCNA) were utilized to search common genes. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis were conducted to explore biological functions. Three machine learning algorithms (least absolute shrinkage and selection operator (LASSO), support vector machine recursive feature elimination (SVM-RFE), and random forest (RF)) were used to further identify shared diagnostic genes, and these genes were assessed via receiver operating characteristic (ROC) curves in discovery and validation datasets. CIBERSORT was employed for immune cell infiltration analysis. Transcription factors (TFs)-genes and miRNAs-genes regulatory networks were conducted by NetworkAnalyst. Finally, relevant drug targets were predicted by DSigDB. Results: Based on DEGs, 173 overlapping genes were obtained and primarily enriched in immune- and inflammation-related pathways. WGCNA revealed 34 common disease-related genes, which were enriched in similar biological pathways. Intersecting the DEGs with WGCNA results yielded 22 candidate genes. Moreover, three machine learning algorithms identified three shared genes (CSF2RB, CXCR4, and LYN) between PD and pSS, and these genes demonstrated good diagnostic performance (AUC>0.85) in both discovery and validation datasets. The immune cell infiltration analysis showed significant dysregulation in several immune cell populations. Regulatory network analysis highlighted that WRNIP1 and has-mir-155-5p might be pivotal co-regulators of the three shared gene expressions. Finally, the top 10 potential gene-targeted drugs were screened. Conclusion: CSF2RB, CXCR4, and LYN may serve as potential biomarkers for the concurrent diagnosis of PD and pSS. Additionally, we identified common molecular mechanisms, TFs, miRNAs, and candidate drugs between PD and pSS, which may provide novel insights and targets for future research on the pathogenesis, diagnosis, and therapy of both diseases.
What problem does this paper attempt to address?