Integrative computational approaches to study protein-nucleic acid interactions

Interactions between proteins and nucleic acid molecules are central to the cellular regulation and homeostasis. To study them, I employ a wide range of computational analysis methods to integrate genomic data from many types of experiment. This thesis has three parts. In the first part, I explore the patterns of indels created by CRISPR-Cas9 genome editing. By thorough characterisation of the precision of editing at thousands of genomic target sites, we identify simple sequence rules that can help predict these outcomes. Furthermore, we examine the role of the structural chromatin context in fine-tuning Cas9-DNA interactions. In the second part, I explore methods to study protein-RNA interactions. I use comparative computational analyses to assess both the data quality of, and data analysis methods for, different crosslinking and immunoprecipitation (CLIP) technologies. I then develop new methods to analyse data generated by hybrid individual-nucleotide resolution CLIP (hiCLIP). By tailoring computational solutions to an understanding of experimental conditions, I improve the overall sensitivity of hiCLIP, and ultimately feedback to drive ongoing experimental development. In the third part, I focus on the Staufen family of double-stranded RNA binding proteins and using hiCLIP data to define transcriptome-wide atlases of RNA duplexes bound by these proteins both in a cell line and in rat brain tissue. Through integration with other data sets, both publicly available and newly generated, I derive insights into their function in RNA metabolism, and in how these interactions change during the course of mammalian brain development with putative roles in ribonucleoprotein complex formation. In summary, I present a range of tailored computational methods and analyses developed to understand interactions between proteins and nucleic acids; aiming to link these interactions to functional outcomes