The Effect Of Nucleosome Positioning On RNA Splicing And The Recognition Of Histone Variants

In eukaryotic genomes, the packaging of DNA around histone octamer not only facilitates the storage and organization of long eukaryotic chromosomes, but also plays a crucial role in diverse DNA-related biological processes, such as transcriptional regulation, replication, repair and recombination. In addition, the substitution between canonical histone and its variant also has important roles in the regulation of gene expression. Revealing genome-wide nucleosome organization and differentiating histone variant containing nucleosomes with canonical nucleosomes will provide novel insights into the full understanding of gene expression.To obtain the accurate information on nucleosome positioning, apart from experimental approaches, the computational methods are also important. With the decoding of the nucleosome positioning code, great progresses have been made in the area of nucleosome positioning prediction. However, except promoter regions, genome-wide properties and functions of nucleosome organization remain ill-determined in mammalian genomes.In this dissertation, by using the method of Increment of Diversity with Quadratic Discriminant (IDQD) analysis, a novel computational model was proposed and applied to analyze nucleosome distributions around splice junctions in constitutive and alternative splicing events for the first time. Based on DNA flexibility, the nucleosome occupancy profiles around splice junctions were also validated by ab initio method. By introducing specific flexibility, RNA sequence flexibility was compared around splice junctions. Finally, H2A.Z and H2A containing nucleosomes were successfully discriminated based on histone methylation. The main contributions are summarized as follows:1. Based on the characteristic of nucleotide distribution in nucleosome positioning and depleted sequences, a novel computational model for nucleosome positioning prediction was proposed by using the method of Increment of Diversity with Quadratic Discriminant (IDQD). The performance of the model was evaluated in two independent datasets. The results show high accuracies, indicating that the model could be employed to annotate genome-wide nucleosome distribution.2. By using structural information of dinucleotide steps, the flexibility of DNA (FD) was defined to describe DNA sequence flexibility. Considering RNA folding free energy, specific flexibility (SF) was defined to describe RNA structural flexibility and employed to predict sequence flexibility of the recently decoded HIV-1 RNA genome. In this procedure, high predictive accuracies were obtained, demonstrating the reasonability of the definition of SF.3. By calculating nucleosome occupancy score (NOScore) and DNA flexibility (FD) for sequences around splice junctions of constitutive, cassette exon, alternative 3'and alternative 5'splicing events in both human and mouse genomes, we found that exons have higher NOScore compared with their flanking intron sequences in both constitutive and alternative splicing events, suggesting the stronger nucleosome occupation potential in exon regions. In addition, the NOScore valleys present at approximately 25 bp upstream of the acceptor site in all splicing events.4. RNA specific flexibility was compared by calculating SF. We found that RNA transcripts from nucleosome occupancy regions are relatively rigid and those from nucleosome depleted regions are relatively flexible. The negative correlation existed between the nucleosome occupation/depletion of DNA sequence and the structural flexibility/rigidity of its transcript around splice junctions may provide clues to the deeper understanding of unexpected roles for nucleosome organization in the regulation of RNA splicing.5. Based on histone methylation and by using IDQD algorithm, H2A.Z and H2A containing nucleosomes were successfully discriminated. By computing DNA sequence flexibility, the average flexibility of DNA wrapped around the H2A.Z containing nucleosome was found to be significantly weaker than that around the H2A containing nucleosome, providing explanations to the different roles of H2A.Z and H2A containing nucleosomes in the process of gene expression.