Gene Expression Patterns in Liquid-Liquid Phase Separation: A Prognostic and Predictive Tool for Colorectal Cancer

To illuminate the role of liquid-liquid phase separation (LLPS) in colorectal cancer (CRC) and its potential as a therapeutic target, we conducted an in-depth analysis focusing on LLPS-related genes (LLPSRGs). Our approach involved a predictive model with LLPSRGs levels a s basis, with intent to prognosticate patient survival and assist in immunotherapy decision-making. Employing univariate Cox regression equipped with LASSO for initial screening, we identified LLPSRGs linked to CRC survival. A subsequent multivariate Cox regression analysis led to the establishment of a prognostic model that included seven pivotal genes. Survival analysis and ROC curve assessment confirmed the accuracy of this model in predicting CRC patient survival, validated in a GEO dataset. PPI network showed that there were complex correlations between differentially expressed genes in risk groups. qRT-PCR findings showed that the top 10 hub genes in the network were all upregulated in CRC cells. Immune profiling unveiled differences in immune cell infiltration and potential immunotherapy responses between high-risk and low-risk groups, with the latter being more suitable for immunotherapeutic interventions. Drug sensitivity analysis supported personalized treatment approaches. Overall, we developed a robust LLPSRG-based prognostic model for CRC, offering insights for CRC clinical prognosis and therapy.