A combination of conserved and stage-specific lncRNA biomarkers to detect lung adenocarcinoma progression.

Lung adenocarcinoma is highly heterogeneous at the molecular level between different stages; therefore, understanding molecular mechanisms contributing to such heterogeneity is needed. In addition, multiple stages of progression are critical factors for lung adenocarcinoma treatment. However, previous studies showed that cancer progression is associated with altered lncRNA expression, highlighting the tissue-specific and developmental stage-specific nature of lncRNAs in various diseases. Therefore, a study using an integrated network approach to explore the role of lncRNA in carcinogenesis was done using expression profiles revealing stage-specific and conserved lncRNA biomarkers in lung adenocarcinoma. We constructed ceRNA networks for each stage of lung adenocarcinoma and analysed them using network topology, differential co-expression network, protein-protein interaction network, functional enrichment, survival analysis, genomic analysis and deep learning to identify potential lncRNA biomarkers. The co-expression networks of healthy and three successive stages of lung adenocarcinoma have shown different network properties. One conserved and four stage-specific lncRNAs are identified as genome regulatory biomarkers. These lncRNAs can successfully identify lung adenocarcinoma and different stages of progression using deep learning. In addition, we identified five mRNAs, four miRNAs and twelve novel carcinogenic interactions associated with the progression of lung adenocarcinoma. These lncRNA biomarkers will provide a novel perspective into the underlying mechanism of adenocarcinoma progression and may be further helpful in early diagnosis, treatment and prognosis of this deadly disease.