Identification of A New Lung Cancer Biomarker Signature using Data Mining and Preliminary In-Vitro Validation

Background: Lung adenocarcinoma is one of the major subtype of non-Small Cell Lung Cancer and biomarkers are essential to be identified for early diagnosis. The study aims to find in silico and preliminary in vitro analysis of potential biomarkers for lung adenocarcinoma. Methods: Bioinformatics analysis in parallel to data mining analysis was performed on microarray data with lung adenocarcinoma samples to identify potent gene biomarkers associated with lung cancer type. Afterwards, these genes were then validated in vitro using RT-qPCR analysis in cancerous (Calu-3) and non-cancerous (MRC-5) cell lines. Moreover, these genes were used in machine learning-based analysis to classify lung adenocarcinoma samples from controls. The analysis includes three experiments—the bioinformatic (in silico), in vitro, and machine learning analyses. Results: The three experiments identified four genes, namely, SLC15A1, GPR123 (ADGRA1), KCNAB2, and KNDC1, as key biomarkers and the most relevant gene features for distinguishing lung adenocarcinoma from control. Conclusions: This study identifies four biomarkers associated with lung adenocarcinoma through bioinformatics, in vitro and machine learning analyses. These four genes shows strong potential for further investigation in clinical research.