Computational analyses of RNA-Sequencing data to identify splicing and polyadenylation regulatory elements

The following dissertation presents my scientific contributions to computational research in the field of RNA biology, carried out in the group of Professor Mihaela Zavolan at the Biozentrum, University of Basel, Switzerland. The projects in which I was involved focused on splicing and polyadenylation, two crucial steps in the maturation of eukaryotic mRNAs, and targets of post-transcriptional mechanisms for regulating gene expression. In the following sections I provide five concise summaries of scientific articles (published, under revision and in preparation) that I co-authored, providing my knowledge and expertise in bioinformatics. Briefly, these manuscripts cover the following aspects. Together with other colleagues from our group we developed a general purpose RNA-seq data processing workflow which utilizes best-practices in scientific software development for reproducible research. I further helped in analyzing the molecular impact of the LARP7 protein on the whole transcriptome, identifying splicing events that are deregulated upon changes in LARP7 expression level. Anal- ogously, I worked with other members of our group in defining targets of the CFIm complex that is involved in alternative polyadenylation and in characterizing the downstream effects on CFIm-dependent polyadenylation on cellular pathways. Further, based on a previously published probabilistic framework for predicting binding sites of nucleic acid-binding proteins, I developed a user-friendly bioinformatics tool to search for binding sites of RNA-binding proteins on mRNAs. Finally, in my main PhD project I combined all the aforementioned expertise together to develop a complex computational workflow to infer the activity of RNA-binding proteins (RBPs) on alternative splicing and alternative polyadenylation from RNA-seq data. The core of my work consisted of a novel computational method to assess the impact of binding sites of RBP regulators on the inclusion of cassette exons in mature transcripts. We validated this method on various RBP knock-down datasets and uncovered proteins which are plausible drivers of differential mRNA processing in glioblastoma cancer. The workflow is implemented in a modular fashion, leaving room for expansion in potential subsequent research projects which are discussed in the final section of this dissertation. In conclusion, I present my novel scientific contribution to understanding the regulatory impact of RNA-binding proteins on mRNA processing from a computational perspective. This work is relevant not only for the results we present but also for creating an opportunity for other scientists to investigate RNA maturation in their research projects. I hope my work will aid others pursuing their scientific passion and that I indirectly contribute to many fascinating discoveries in the future.