Abstract
OBJECTIVE: Cuproptosis-related genes are closely related to lung adenocarcinoma (LUAD), which can be analyzed via the analysis of long noncoding RNA (lncRNA). To date, the clinical significance and function of cuproptosis-related lncRNAs are still not well elucidated. Further analysis of cuproptosis-related prognostic lncRNAs is of great significance for the treatment, diagnosis, and prognosis of LUAD. METHODS: In this study, a multiple machine learning (ML)-based computational approach was proposed for the identification of the cuproptosis-related lncRNAs signature (CRlncSig) via comprehensive analysis of cuproptosis, lncRNAs, and clinical characteristics. The proposed approach integrated multiple ML algorithms (least absolute shrinkage and selection operator regression analysis, univariate and multivariate Cox regression) to effectively identify the CRlncSig. RESULTS: Based on the proposed approach, the CRlncSig was identified from the 3450 cuproptosis-related lncRNAs, which consist of 13 lncRNAs (CDKN2A-DT, FAM66C, FAM83A-AS1, AL359232.1, FRMD6-AS1, AC027237.4, AC023090.1, AL157888.1, AL627443.3, AC026355.2, AC008957.1, AP000346.1, and GLIS2-AS1). CONCLUSIONS: The CRlncSig could well predict the prognosis of different LUAD patients, which is different from other clinical features. Moreover, the CRlncSig was proved to be an effective indicator of patient survival via functional characterization analysis, which is relevant to cancer progression and immune infiltration. Furthermore, the results of RT-PCR assay indicated that the expression level of FAM83A-AS1 and AC026355.2 in A549 and H1975 cells (LUAD) was significantly higher than that in BEAS-2B cells (normal lung epithelial).