Related Experiment Video
Updated: Jan 17, 2026
Implementation of In Vitro Drug Resistance Assays: Maximizing the Potential for Uncovering Clinically Relevant Resistance Mechanisms
Published on: December 9, 2015
Integrative single-cell and machine learning analysis predicts lactylation-driven therapy resistance in prostate
Zhiyu Liu1,2, Yuqi Li1,2, Juan Wang1
1Department of Urology, Affiliated Hospital of Southwest Medical University, Luzhou, Sichuan, China.
Background:
Prostate cancer (PCa) is a common malignancy in males. Predicting its prognosis and addressing drug resistance remain challenging. This study develops a novel prognostic model focusing on lactylation and resistance, which plays a crucial role in tumor biology.
Methods:
Single-cell analysis was employed to identify subpopulations expressing lactylation-related genes. Transcriptomic sequencing was used to identify drug resistance-associated genes. Univariate Cox proportional hazards models and machine learning techniques were used to identify prognostic genes, assisting in the development of a risk assessment framework. Additionally, we investigated how features related to lactylation and drug resistance correlate with clinical characteristics, the tumor microenvironment, and treatment responses, revealing potential interconnections.
Results:
In this study, a model composed of 29 biomarkers was developed by integrating single-cell data and machine learning algorithms. The model predictive efficacy was validated through Kaplan-Meier (KM) analysis, univariate Cox (HR=3.59, 95%CI: 2.78-4.63) and multivariate Cox (HR=2.81, 95%CI: 1.96-4.03) regression. Comprehensive analysis revealed significant differences in tumor immune dysfunction and exclusion (TIDE) scores, immunophenoscore (IPS) scores, and chemotherapy drug sensitivity between high-risk and low-risk groups, suggesting that specific biomarkers may be closely associated with prognosis. Furthermore, molecular docking analysis and experiments were conducted to explore the relationship between drug resistance and risk gene-encoded proteins.
Conclusions:
The prognostic model effectively predicts the progression-free interval (PFI) and drug response, with accurate risk stratification for PCa patients. Our findings highlight the potential of risk genes in the development of personalized treatment strategies and enhancing PCa prognostic assessment.
