Construction of a bioinformatics-based prognostic model related to liquid-liquid phase separation in lung adenocarcinoma

Abstract

Abstract It was reported that liquid-liquid phase separation (LLPS) related genes (LRGs) were involved in the prognosis of a variety of tumors. We aimed to investigate the impact of LRGs on lung adenocarcinoma (LUAD) patients' prognosis, providing further insights for patient treatment and prognosis. TCGA-LUAD, GSE31210 and GSE131907 were applied in this study. Differentially expressed LRGs (DE-LRGs) were identified by intersecting the differentially expressed genes (DEGs) obtained through differential expression analysis with the LRGs acquired from the online database. A total of 17 DE-LRGs were gained by intersecting 5,445 DEGs and LRGs obtained from the online database. Subsequently, TACC3, TPX2, PRC1, FGFR2, ORC1, and PLK4 were identified as prognostic genes for constructing a risk model. The high-risk patients exhibited shorter survival time compared to the low-risk patients in both TCGA-LUAD and GSE31210. The nomogram of pathologic stage and riskScore demonstrated good predictive ability for the 1/3/5-year survival rate of LUAD patients. Next, immune analysis indicated TPX2 exhibited the strongly positive correlation with M0 and M1 macrophages, which displayed the important role of macrophages in the phase separation mechanism. Single cell analysis revealed higher expression levels of prognostic genes in T lymphocytes. We speculated that prognostic genes contributed to tumor development by affecting T-cell activity. Eventually, we verified the differential expression of prognostic genes by testing the expression of clinical samples. In conclusion, TACC3, PRC1, ORC1, and PLK4 were identified to be associated with LLPS on LUAD in this study. It provided further insights into the treatment of LUAD.