Analysis and validation of novel biomarkers related to palmitoylation in adenomyosis.

Background: Adenomyosis, a common gynecological disorder in women of reproductive age, is characterized by endometrial invasion into the myometrium, leading to uterine enlargement and smooth muscle hypertrophy. Typical clinical symptoms include chronic pelvic pain, abnormal uterine bleeding, and infertility, which significantly impair patients' quality of life. Currently, effective diagnostic biomarkers for adenomyosis are lacking. Recent studies suggest that estrogen may promote Scribble protein depalmitoylation by upregulating APT1 and APT2 expression. Depalmitoylation facilitates Scribble's translocation from the cell membrane to the cytoplasm, disrupting endometrial epithelial cell polarity. This polarity loss may enhance abnormal proliferation, migration, and invasion of endometrial epithelial cells, promoting endometrial tissue infiltration into the myometrium and contributing to adenomyosis development and progression. Therefore, investigating adenomyosis diagnosis and treatment from the perspective of palmitoylation-related genes holds significant scientific importance. Methods: In this study, four datasets, GSE244236, GSE190580, GSE185392 and GSE157718, were downloaded and the data were screened and standardized the data. First, GSE244236 was used as the training dataset. By integrating multiple bioinformatics approaches-including differential gene analysis (DEGs), weighted gene co-expression network analysis (WGCNA), Least Absolute Shrinkage (LASSO), random forest (RF) methods, and Support Vector Machine-recursive feature elimination (SVM-RFE)-we identified three overlapping diagnostic genes through comprehensive analysis. Meanwhile, the diagnostic value of each biomarker was assessed using the receiver operating characteristic curve analysis in the remaining three datasets. In addition, single-sample gene set enrichment analysis (ssGSEA) were utilized to explore the infiltration of immune cells in adenomyosis and to examine the correlation between diagnostic biomarkers and immune cells. Results: A total of 549 differentially expressed genes were identified in the analysis. Through WGCNA analysis, we obtained 25 palmitoylation-related intersecting genes. Using LASSO, RF and SVM-RFE algorithms, seven potential diagnostic genes were finally screened: LIPH, CYP2E1 and CHRNE. Conclusion: In this study, we successfully identified diagnostic biomarkers for adenomyosis using comprehensive bioinformatics analysis and machine learning methods, and validated them with nomogram and ROC curves. Our findings provide new perspectives for understanding the pathogenesis of palmitoylation-related genes in adenomyosis and offer potential targets for the development of new therapeutic strategies.