Gene Expression Omnibus Database

Gene expression data of paired NSCLC patients’ cancer tissues (tumor tissues and para-cancerous tissues) and healthy human serum were obtained from Gene Expression Omnibus (GEO)¹ database, and screened for differentially expressed miRNAs with a p-value less than 0.05. Studies were regarded eligible for our analysis according to the following criteria: (1) Studies with non-small cell lung cancer tissue or serum samples, (2) Studies with information about the technology utilized, and (3) Studies with the existence of normal cohorts as the control. Based on these rules, five datasets were obtained from the GEO database, including tumor tissues from non-small lung cancer patients and adjacent tissues ({"type":"entrez-geo","attrs":{"text":"GSE2109","term_id":"2109"}}GSE2109, {"type":"entrez-geo","attrs":{"text":"GSE74190","term_id":"74190"}}GSE74190, and {"type":"entrez-geo","attrs":{"text":"GSE63805","term_id":"63805"}}GSE63805), and serum from lung cancer patients and healthy controls ({"type":"entrez-geo","attrs":{"text":"GSE64591","term_id":"64591"}}GSE64591 and {"type":"entrez-geo","attrs":{"text":"GSE20189","term_id":"20189"}}GSE20189). Next, we performed the differential analysis (|Log₂FC| > 2, p-value < 0.05) by comparing tumor tissues to normal tissues in the SPSS 19.0 software using student’s t-test. Subsequently, we combined the differentially expressed genes acquired from GEO and gene chips to get the shared gene signatures.