Computational Methods for High-Throughput Small RNA Analysis in Plants

In plants, millions of uncharacterized small RNAs (sRNAs) can be found for which experimental validation poses an impractical laborious task. On the other hand, it is nowadays relatively easy and cheap to capture sRNA sequences at a genome-wide scale using deep sequencing approaches. Computational methods have been devised to perform preliminary studies of populations of sRNAs and guide downstream experiments. Nonetheless, sRNA biology is complex and demands a battalion of independent computational methods which currently cannot be found in one unifying framework necessary for a thorough examination, while many critical algorithms are inaccurate or remain to be devised. In special, sRNAs that guide epigenetic mechanisms such as DNA methylation are estimated to be among the most abundant in plants, but despite the importance of this category, there is a lack of tools available in the public domain for their identification.
In this thesis, novel computational methods for sRNA categorization are introduced, as well as an integrative framework for general sRNA analysis. The new software, developed for high-throughput sRNA analysis, was applied to real problems in biology bringing new insights to sRNA-mediated epigenetics.