Integrative machine learning-driven prognosis and immunotherapy stratification via lactylation-associated gene in ovarian cancer.

This study utilized a comprehensive approach by integrating multi-omics data to systematically assess lactate modification levels across diverse cell types employing AUCell, JASMINE, and singscore algorithms. An epithelial subpopulation exhibiting the highest lactylation score was successfully pinpointed, and differentially expressed genes linked to lactylation were identified. Through machine learning techniques, a prognostic model was developed based on three genes (TMEM126B, PYGL, and NDUFS6). This model displayed significant associations with immune tumor microenvironment characteristics, microsatellite instability, immune checkpoint expression, and tumor mutation burden. Elevated lactylation risk was linked to the activation of cell cycle and oncogenic pathways, dampened anti-tumor immune responses, and increased expression of immune checkpoints, indicating potential limitations in immunotherapy efficacy. Noteworthy, NDUFS6 exhibited significant upregulation in ovarian cancer (OC) tissues and correlated with an unfavorable prognosis. Functional investigations demonstrated that NDUFS6 knockdown suppressed OC cell proliferation and induced cell cycle arrest. Remarkably, D-lactose emerged as a promising therapeutic agent targeting NDUFS6, underscoring its potential for precise OC treatment.