UC San Diego

This dissertation describes an integration of computational and biological techniques to characterize small RNA genes in two key marine organisms, a sea squirt and a diatom. Eukaryotic small RNA genes, which are typically 19-31 nucleotides in length, regulate gene expression within cells in a temporal and state-dependent manner, controlling essential processes such as embryological development, cell differentiation, responses to environmental stress, and cellular defense mechanisms. In the first chapter, computational methods were developed to predict evolutionarily conserved members of one class of small RNAs, known as microRNAs, in the sea squirt, Ciona intestinalis. The sea squirt is an important model organism due to its phylogenetic placement at the emergence of vertebrates. The microRNA prediction algorithm was designed to quickly screen the genome for the presence of conserved microRNAs, producing the first validated collection of microRNAs in the sea squirt. Additionally, a target prediction algorithm was implemented which identified potential target genes. The second chapter built upon these techniques and expanded the search for other classes of endogenous small RNAs in the diatom, Thalassiosira pseudonana. Diatoms are unicellular phytoplankton, chosen as the model organism because of their global importance in processes such as carbon fixation and nutrient cycling. A small RNA cDNA library was constructed for exponentially growing T. pseudonana, and then pyrosequenced. Computational analysis of approximately 300,000 sequences yielded strong evidence of small RNA genes in T. pseudonana, including microRNAs, repeat-associated short interfering RNAs, and endogenous short interfering RNAs. The third chapter focused on differential small RNA gene expression in Thalassiosira pseudonana, under various nutrient stress conditions. Small RNA cDNA libraries were constructed under conditions of exponential growth, silicon starvation, nitrogen starvation, and iron starvation. The libraries were then processed with high throughput SOLiD sequencing. A methodology was developed to computationally analyze the 150 million sequences, generating a profile of differential small RNA expression between the conditions, as well as a core subset of small RNAs expressed across all conditions. The novel computational techniques implemented in this dissertation can be applied to other organisms and aid in elucidating the roles of small RNAs in gene regulation