Identification of four differentially methylated genes as prognostic signatures for stage I lung adenocarcinoma

Abstract

Background: Lung adenocarcinoma (LUAD) is the main subtype of non-small cell lung cancer with a low survival prognosis. We aimed to generate a prognostic model for the postoperative recurrence of LUAD.

Methods: The methylated DNA data of LUAD patients were downloaded from the Cancer Genome Atlas (TCGA). The differentially methylated genes were identified and protein-protein interacting network was constructed, with which prognostic signature of this cancer was generated. Survival and functional pathways analysis w used to evaluate the clustering ability of the prognostic signature.

Results: We identified 151 differentially methylated genes related to relapse-free survival of patients with LUAD. Nine hub genes were identified in PPI network, with which 4 gene pair signature was selected as prognostic signature. The potential functions of 6 genes (JDP2, SERPINA5, PLG, SEMG2, RFX5, and POLR3B) in the 4-gene pair signature were enriched in intracellular protein synthesis and transportation.

Conclusion: The four gene pair signature can predict the prognosis of patients with stage I LUAD. Our study provides a reference for patients with postoperative adjuvant therapy.

Keywords: DNA methylation; Lung adenocarcinoma; Prognostic signature; Survival analysis.

Fig. 1

PPI network based on the survival 151 genes. Node represents the gene in the PPI network. Red represents the degree of the gene is over 3; blue represents the degree of the gene less than 3

Fig. 2

The association between four gene-pair signature and the RFS of 191 LUAD patients. The log rank p value is displayed separately. a High- and low-risk group clustered by gene pair 1; b high- and low-risk group clustered by gene pair 2; c high- and low-risk group clustered by gene pair 3; d high- and low-risk group clustered by gene pair 4

Fig. 3

Survival analysis of 4 gene-pair signature. This signature was able to classify the stage I LUAD patients into two groups with significant RFS (univariate COX p < 0.0057, multivariate COX p < 0.0069). a Multivariate survival analysis of the signature; b univariate survival analysis of the signature; c survival analysis on the impacts of gender; d survival analysis on the impacts of smoking; e survival analysis on the impacts of age. The log rank p value is displayed separately