关键词: 多囊卵巢综合征, 差异表达基因, 微阵列数据, eQTL分析, 孟德尔随机化, 免疫细胞浸润

Abstract:

Objective To identify novel genetic targets for polycystic ovary syndrome (PCOS) and to reveal the molecular basis underlying the development of PCOS using microarray datasets. Methods A comprehensive analysis was conducted on three independent PCOS datasets from the Gene Expression Omnibus database, and data processing and normalization were performed using R software. The evaluation of the relationship between differentially expressed genes (DEGs) and PCOS included differential expression analysis, expression quantitative trait locus analysis, and Mendelian randomization (MR) analysis. Gene Set Variation Analysis and Gene Ontology/Kyoto Encyclopedia of Genes and Genomes enrichment analysis were performed to explore the functional roles and pathways of these genes. Lasso regression, SVM machine learning, and random forest tree method were used to screen PCOS intersection feature genes, and then a cross validation was conducted with MR method. In addition, CIBERSORT analysis was used to evaluate the infiltration levels of 22 immune cells in PCOS. Finally, a single gene ceRNA regulatory network was constructed for visualization. Results This study identified 59 significant differentially expressed genes (DEGs), including 28 upregulated DEGs and 31 downregulated DEGs. Lasso regression, SVM machine learning, and random forest tree method were used to screen six intersection feature genes related to PCOS, including DNAJC3, ASPH, TLR4, SEC24D, SGK1 and AMFR. These genes were involved in important biological processes and pathways, including responses to endoplasmic reticulum stress, chemical stress, autophagy regulation, endogenous apoptosis, lipid and atherosclerosis, neurodegeneration, etc. The MR analysis method for the cross validation of the aforementioned three methods further suggested that SEC24D had a causal relationship with PCOS, and 37 miRNAs and 15 lncRNAs were closely related to SEC24D. CIBERSORT analysis showed a unique distribution of immune cells in PCOS, particularly a significant increase in neutrophil count, emphasizing the importance of immune processes in PCOS. Conclusion This study provides new insights into the molecular basis of PCOS and emphasizes the potential for therapeutic interventions and targeted molecular pathways for treating PCOS, laying the foundation for further research and clinical work.

Key words: polycystic ovary syndrome, differentially expressed genes, microarray data, eQTL analysis, Mendelian randomization, immune cell infiltration