Computational analysis of chromatin features associated with transcription and recombination in yeast

• Main Author: Hansen, Loren
• Language: en_US
• Published: Boston University 2018
• Online Access: https://hdl.handle.net/2144/31563

Thesis (Ph.D.)--Boston University === PLEASE NOTE: Boston University Libraries did not receive an Authorization To Manage form for this thesis or dissertation. It is therefore not openly accessible, though it may be available by request. If you are the author or principal advisor of this work and would like to request open access for it, please contact us at [e-mail]. Thank you. === When and where transcription factors (TFs) bind is of primary importance in determining which genes are expressed. Many transcriptional regulators bind specific DNA motifs. These motifs are often short and degenerate with only a small subset of motifs serving as binding platforms. By occupying potential binding sites nucleosomes can help determine which of the pool of motifs found in the genome are available for binding. This poses the question of how nucleosome positioning is determined such that incorrect motifs are covered while correct motifs are available for binding. Here I show that many motifs recognized by TFs show strong positional preferences to occur predominantly in potential regulatory regions. I also demonstrate that TFs whose motifs show the strongest positional preference have a tendency to possess chromatin modifying properties. By altering the local chromatin structure remodeling factors whose motifs show strong positional preferences can assist in determining DNA binding specificity for other TFs. To further pursue the question of TF binding specificity it would be helpful to identify differences in the local genomic environment surrounding bound motifs compared to unbound motifs. To this end, I obtained a set of motifs bound by protein and a set of motifs not bound by protein for 9 TFs. For each motif a vector of biological features was constructed representing the local genomic environment around that motif. Using machine learning based approaches a number of striking differences in the genomic environment were identified between bound and unbound motifs. Meiotic recombination is not distributed uniformly throughout the genome. There are regions of high and low recombination. Understanding what factors regulate the frequency of recombination is an area of active research. I describe a method to predict biological features associated with recombination using a feature selection-based approach. Some of my predictions have previously been shown to be associated with recombination. A number of my predictions are novel and represent promising candidates for further study. I show that nucleosome occupancy maps produced using next generation sequencing exhibit a bias at recombination hot spots and this bias is strong enough to obscure biologically relevant information. === 2031-01-01