Premium
A prognostic 4‐gene expression signature for squamous cell lung carcinoma
Author(s) -
Li Jun,
Wang Jing,
Chen Yanbin,
Yang Lijie,
Chen Sheng
Publication year - 2017
Publication title -
journal of cellular physiology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.529
H-Index - 174
eISSN - 1097-4652
pISSN - 0021-9541
DOI - 10.1002/jcp.25846
Subject(s) - gene , biology , proportional hazards model , cancer research , genetics , medicine
Squamous cell lung carcinoma (SQCLC), a common and fatal subtype of lung cancer, caused lots of mortalities and showed different outcomes in prognosis. This study was to screen key genes and to figure a prognostic signature to cluster the patients with SQCLC. RNA‐Seq data from 550 patients with SQCLC were downloaded from The Cancer Genome Atlas (TCGA). Genetically changed genes were identified and analyzed in univariate survival analysis. Genes significantly influencing prognosis were selected with frequency higher than 100 in lasso regression. Meanwhile, area under the curve (AUC) values and hazard ratios (HR) for seed genes were obtained with R Language. Functional enrichment analysis was performed and clustering effectiveness of the selected common gene set was analyzed with Kaplan–Meier. Finally, the stability and validity of the optimal clustering model were verified. A total of 7,222 genetically changed genes were screened, including 1,045 ones with p < 0.05, 1,746, p < 0.1, and 2,758, p < 0.2. The common gene sets with more than 100 frequencies were 14‐Genes, 10‐Genes and 6‐Genes. Genes with p < 0.05 participated in positive regulation of ERK1 and ERK2 cascade, angiogenesis, platelet degranulation, cell–matrix adhesion, extracellular matrix organization, macrophage activation, and so on. A four‐gene clustering model in 14‐Genes ( DPPA , TTTY16 , TRIM58 , HKDC1 , ZNF589 , ALDH7A1 , LINC01426 , IL19 , LOC101928358 , TMEM92 , HRASLS , JPH1 , LOC100288778 , GCGR ) was verified as the optimal. The discovery of four‐gene clustering model in 14‐Genes can cluster the patient samples effectively. This model would help predict the outcomes of patients with SQCLC then improve the treatment strategies.