TY - JOUR AU - Wu, Chi-Yun AU - Chuang, Eric Y. AU - Lu, Tzu-Pin PY - 2016 TI - Low correlation of lncRNA and target gene expression in microarray data JF - Translational Cancer Research; Vol 5, No 2 (April 22, 2016): Translational Cancer Research Y2 - 2016 KW - N2 - Background: Recent studies have indicated that long non-coding RNAs (lncRNAs) play important roles in regulating the expression levels of genes and proteins. Challenges arise when trying to identify the target genes of lncRNAs. A popular approach is to select possible target genes of specific lncRNAs based on the similarity of their expression patterns. However, such associations have not been validated in the context of the whole human genome. Methods: To address this issue, ten microarray datasets with at least 100 samples from five tissue types were analyzed in this study. All datasets were examined using the Affymetrix u133plus 2.0 platform. Probes targeting lncRNAs were identified by performing re-annotation of the probe sequences, and the link between lncRNAs and their target genes was retrieved from an online database. We manually annotated the link file with information about the regulation mechanism and tissue types. Results: A total of 956 lncRNAs and their target genes were identified as having regulation at the transcriptional level. Pearson correlation coefficients were calculated in all datasets, and low correlations were observed. A resampling test demonstrated that the expression levels of lncRNAs and their target genes showed similar correlations in the pairs obtained by real data and random selection. Further investigations were performed by identifying lncRNAs with differential expression and tissue specificity; however, low correlations were still observed. Conclusions: In conclusion, gene expression microarray data are not a valid way to identify possible target genes of lncRNAs. UR - https://tcr.amegroups.org/article/view/7350