Abstract
Noncoding RNA play important roles in various biological processes and diseases, including cancer. The expression profile of circular RNA (circRNA) has not been systematically investigated in lung adenocarcinoma (LUAD). In this study, we performed genomewide transcriptome profiling of coding genes, long noncoding RNA (lncRNA), and circRNA in paired LUAD and nontumor tissues by ribosomal RNA-depleted RNA sequencing. The detected reads were first mapped to the human genome to analyze expression of coding genes and lncRNA, while the unmapped reads were subjected to a circRNA prediction algorithm to identify circRNA candidates. We identified 1282 differentially expressed coding genes in LUAD. Expression of 19 023 lncRNA was detected, of which 244 lncRNAs were differentially expressed in LUAD. AFAP1-AS1, BLACAT1, LOC101928245, and FENDRR were most differentially expressed lncRNAs in LUAD. Also identified were 9340 circRNA candidates with ≥ 2 backspliced, including 3590 novel circRNA transcripts. The median length of circRNA was ~ 530 nt. CircRNA are often of low abundance, and more than half of circRNAs we identified had < 10 reads. Agarose electrophoresis and Sanger sequencing were used to confirm that four candidate circRNA were truly circular. Our results characterized the expression profile of coding genes, lncRNA, and circRNA in LUAD; 9340 circRNAs were detected, demonstrating that circRNA are widely expressed in LUAD. DATABASE: The raw RNA sequencing data have been submitted to Gene Expression Omnibus (GEO) database and can be accessed with the ID GEO: http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE104854.