Discovery of barley miRNAs through deep sequencing of short reads

Abstract

Our study, the first large scale investigation of small RNAs in barley, has identified up to 100 miRNAs. We demonstrate that reliable identification of miRNAs via deep sequencing in a species whose genome has not been sequenced requires a more careful analysis of sequencing errors than is commonly performed. We devised a read filtering procedure for dealing with errors. In addition, we found that the use of a large dataset of almost 35 million reads permits the use of read abundance distributions along putative precursor sequences as a practical tool for isolating miRNAs in a large background of reads originating from other non-coding and coding RNAs. This study therefore provides a generic approach for discovering novel miRNAs where no genome sequence is available.