On the impact of batch effect correction in TCGA isomiR expression data. Issue 1 (11th March 2021)

• Record Type: Journal Article
• Authors: Ibing, Susanne; Michels, Birgitta E; Mosdzien, Moritz; Meyer, Helen R; Feuerbach, Lars; Körner, Cindy
• Is Part Of: NAR cancer. Volumer 3:Issue 1(2021)
• Publication Date: 2021-03-11
• Subjects: Cancer -- Periodicals; Cancer -- Genetic aspects -- Periodicals; Nucleic acids -- Periodicals; Molecular biology -- Periodicals; 616.994
• DOI: 10.1093/narcan/zcab007
• Languages: English
• ISSNs: 2632-8674

Abstract

Abstract: MicroRNAs (miRNAs) are small non-coding RNAs with diverse functions in post-transcriptional regulation of gene expression. Sequence and length variants of miRNAs are called isomiRs and can exert different functions compared to their canonical counterparts. The Cancer Genome Atlas (TCGA) provides isomiR-level expression data for patients of various cancer entities collected in a multi-center approach over several years. However, the impact of batch effects within individual cohorts has not been systematically investigated and corrected for before. Therefore, the aim of this study was to identify relevant cohort-specific batch variables and generate batch-corrected isomiR expression data for 16 TCGA cohorts. The main batch variables included sequencing platform, plate, sample purity and sequencing depth. Platform bias was related to certain length and sequence features of individual recurrently affected isomiRs. Furthermore, significant downregulation of reported tumor suppressive isomiRs in lung tumor tissue compared to normal samples was only observed after batch correction, highlighting the importance of working with corrected data. Batch-corrected datasets for all cohorts including quality control are provided as supplement. In summary, this study reveals that batch effects present in the TCGA dataset might mask biologically relevant effects and provides a valuable resource for research on isomiRs in cancer (accessible through GEO: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE164767 ). Graphical Abstract: