Identification of reference miRNAs in human tumors by TCGA miRNA-seq data.

Although the accuracy of detecting the expression of miRNAs by quantitative real-time polymerase chain reaction (qRT-PCR) is highly dependent on reliable reference miRNAs, many commonly used reference miRNAs are not stably expressed and as such are not suitable for quantification and normalization of qRT-PCR data. To solve this problem, we analyzed the global expression profiles of thousands of samples in 14 types of common human tumors released by The Cancer Genome Atlas (TCGA), and identified the most stably and highly expressed miRNAs as candidate reference miRNAs in each type of tumor. We found that miR-361-5p and let-7i-5p were the most recommended candidate reference miRNAs in nine and eight types of tumors, respectively, followed by let-7a-5p, mir-28-5p and miR-99b-5p. Our results are of important value to those researchers focused on miRNA; however, these candidate reference miRNAs still need to be validated prior to their use in qRT-PCR studies.