Construction of a Prognostic DNA Damage Repair Gene-related LncRNA Signature by Integrated Analysis of Bulk- and Single-cell RNA Sequencing Data in Lung Adenocarcinoma

Abstract

Despite the development of technology, the prognosis of patients with lung adenocarcinoma (LUAD) has not improved. Therefore, we sought to investigate the potential clinical utility of a risk prognosis model in prognostic stratification. For this purpose, DNA damage repair gene-related (DDRG-related) long noncoding RNAs (lncRNAs) were screened based on a single-cell RNA transcriptome to construct such a model. A total of 510 LUAD samples were selected from The Cancer Genome Atlas-LUAD (TCGA-LUAD) dataset. Samples were divided into two immune subtypes (S1 and S2) after calculating the stromal score, immune score, tumor purity, and immune infiltration in each TCGA-LUAD cohort based on the Estimation of Stromal and Immune cells in MAlignant Tumour tissues using Expression data (ESTIMATE) and the NbClust package. Subsequently, DDRG-related lncRNAs were selected by single-cell data analysis combined with bulk sequencing. Next, DDRG-related lncRNAs were screened through the least absolute shrinkage and selection operator, as well as univariate and multivariate Cox regression analyses to develop a precise DDRG-related-lncRNA prognosis risk model. The functions of the target genes of these lncRNAs were described by the Gene Ontology (GO) enrichment analysis. The prognostic capability of the model was tested by analyzing the expression data of LUAD samples downloaded from the Gene Expression Omnibus database. The correlations between tumor mutational burden, N6-methyladenosine (m6A) gene expression, risk score, and immune score were analyzed. A seven DDRG-related lncRNAs risk prognosis model was established. Based on the model, the TCGA-LUAD cohorts and testing sets were classified into low- and high-risk groups. The former group had better overall survival versus the latter group. In addition, a nomogram based on the risk score and clinical stage showed good calibration and moderate discriminative ability. Using single-cell transcriptome datasets, we constructed and verified a useful DDRG-related-lncRNA risk prognosis model for prognostic stratification in LUAD.