Prognostic evaluation of the novel blueprint of DNA methylation sites by integrating bulk RNA-sequencing and methylation modification data in endometrial cancer.

INTRODUCTION: Endometrial cancer (EC) is a prevalent malignancy affecting the female population, with an increasing incidence among younger age groups. DNA methylation, a common epigenetic modification, is well-established to play a key role in cancer progression. We suspected whether DNA methylation could be used as biomarkers for EC prognosis. METHODS: In the present study, we analyzed bulk RNA-sequencing data from 544 EC patients and DNA methylation data from 430 EC patients in the TCGA-UCEC cohort. We applied weighted correlation network analysis to select a key gene set associated with panoptosis. We conducted correlation analysis between transcriptomic data of the selected key genes and DNA methylation data to identify valuable DNA methylation sites. These sites were further screened by Cox regression and least absolute shrinkage and selection operator analysis. Immune microenvironment differences between high-risk and low-risk groups were assessed using single-sample gene set enrichment analysi, xCell and MCPcounter algorithms. RESULTS: Our results identified five DNA methylation sites (cg03906681, cg04549977, cg06029846, cg10043253 and cg15658376) with significant prognostic value in EC. We constructed a prognostic model using these sites, demonstrating satisfactory predictive performance. The low-risk group showed higher immune cell infiltration. Notably, methylation of site cg03906681 was negatively related to CD8 T cell infiltration, whereas cg04549977 exhibited positive correlations with immune infiltration, particularly in macrophages, activated B cells, dendritic cells and myeloid-derived suppressor cells. PD0325901_1060 was strongly correlated with risk scores, indicating a potential therapeutic response for high-risk EC patients. CONCLUSION: We have developed a robust DNA methylation-based prognostic model for EC, which holds promise for improving prognosis prediction and personalized treatment approaches. These findings may contribute to better management of EC patients, particularly in identifying those at higher risk who may benefit from tailored interventions.

Metabolite fingerprinting of the ripening process in Pixian douban using a feature-based molecular network and metabolomics analysis.

The unique flavor of Pixian douban (PXDB) is widely acknowledged to be associated with its maturation process. However, there is limited knowledge about the non-volatile metabolites that contribute to this flavor. To bridge this gap, this study employed a metabolomics approach and a feature-based molecular network (FBMN) analysis to investigate the non-volatile metabolite fingerprints of PXDB during its two-year maturation process. Specifically, the FBMN tool was utilized to annotate the flavonoid, amide derivatives, and lipid components of PXDB for the first time. Subsequently, the MolNetEnhancer tool was employed to complement the FBMN annotation and identify eight substructural components. Finally, metabolomics analysis was carried out to identify 45 key metabolites involved in flavor formation across 10 major metabolic pathways (p < 0.05). Overall, the findings of this study have significantly expanded our understanding of the non-volatile metabolite fingerprinting and flavor formation mechanisms.