Lessons from microRNA Sequencing Using Illumina Technology

Abstract

The numbers of reads generated by second-generation sequencing technologies permit to establish in a single sequencing lane multiple microRNA (miRNA) expression profiles from small RNA-derived cDNA libraries tagged by barcodes consisting of few bases. Multiplex sequencing allows sample size expansion and thus the statistical reliability of generated data. This allows the detection of discrete changes in miRNA expression levels that occur at the onset of cellular processes. With the development of the “by-amplification” strategy, tagging cDNA libraries is no more a source of technical variability. However, other specific features should be kept in mind when designing experiments aimed at profiling miRNA expression using Illumina sequencing technology, the most important being the substantial distortion between miRNA expression in sequencing data and the true miRNA abundancy. miRNAs of low expression in profiles may correspond to abundant miRNAs in samples and vice versa. We report here data obtained from rat cerebellum and liver that illustrate 1) the high 3’ adaptor dependency of miRNA expression profiles, 2) the impact of sample size when working with moderate (3 - 4 fold) changes of miRNA expression and 3) the impact of the statistical tools used to identify differentially expressed miRNAs.