Search

The advent of high-throughput RNA sequencing (RNA-seq) methods have made it possible to sequence transcriptomes for the cell-wide identi�cation of small non-coding RNAs (sRNAs) and to assess their regulation using di�erential expression analysis by comparing two or more di�erent conditions. During an analysis of a typical set of sRNA sequencing (sRNA-seq) libraries, a large variety of tools and methods are used on the dataset in order to understand the data's quality, content, and to summarise the knowledge gained from the entire analysis. Many of the tools available to do this were created for mRNA sequencing (mRNA-seq) datasets. In this thesis, we present and implement a processing pipeline that can be used to assess the quality and the di�erential expression of sRNA-seq datasets over two or more di�erent conditions. We then utilise aspects of this pipeline in various sRNA-seq experiments. Firstly, we combine our pipeline with current tools for miRNA identi�cation to assess the regulation of miRNAs during larval caste di�erentiation in a novel genome; the European bumblebee (Bombus terrestris). Secondly, we explore the di�erential expression during cell stress of all classes of sRNAs using two cell lines in humans. We also �nd that a speci�c protein, Ro60, is required for the expression of mRNA-derived sRNAs during stress, similar to the way in which sRNAs derived from Y RNAs are regulated. Finally, we utilise our understanding of sRNA mapping patterns, alongside current tools for miRNA identi�cation, to search for functional miRNAs and other sRNAs in the novel genomes of two diatoms. The lack of canonical miRNA predictions in this study has repercussions for the evolutionary theory behind miRNAs. The implementation of our pipeline for sRNA-seq data provides an interactive and quality controlled work ow that can be used to process a dataset from raw sequences to the results of several di�erential expression experiments for all identi�ed sRNA classes within a sequenced transcriptome.