Identification of common diagnostic genes and molecular pathways in endometriosis and systemic lupus erythematosus by machine learning approach and in vitro experiment.

Growing research suggests that endometriosis and systemic lupus erythematosus (SLE) are both chronic inflammatory diseases and closely related, but no studies have explored their common molecular characteristics and underlying mechanisms. Based on GEO datasets, differentially expressed genes in the endometriosis cohort and the SLE cohort were screened using Limma and weighted gene co-expression network analysis (WGCNA), and prediction signatures were constructed using LASSO logistic regression analysis, respectively. Four co-diagnostic genes (PMP22, QSOX1, REV3L, SP110) were identified for endometriosis and SLE. The nomogram, calibration curve, decision curve analyses (DCA), area under the receiver operating characteristic (AUC) curve and external datasets were used to evaluate the diagnostic and predictive value of co-diagnostic genes. The AUC value of the four co-diagnostic genes were higher than 0.85 in both endometriosis and SLE cohorts. Besides, functional enrichment analysis showed that DNA replication, base excision repair, cell cycle and cell adhesion molecules were significantly enriched. Multifactor regulatory network of four co-diagnostic genes was constructed including 96 TFs, 42 miRNA, 43 lncRNA, and 189 drugs, and Tributyrin was found to act on four co-diagnostic genes simultaneously. We identified and validated four co-diagnostic genes and revealed the potential molecular mechanisms of endometriosis and SLE, which is helpful for early diagnosis and targeted therapy. Our study provides a novel perspective for individualized treatment of patients with endometriosis and SLE.