Chromatin Topology and Transcription in Myogenesis

Citation

Fisher-Aylor, Katherine Irene (2017) Chromatin Topology and Transcription in Myogenesis. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/Z9K07290. https://resolver.caltech.edu/CaltechTHESIS:09302016-110627143

Abstract

High-throughput sequencing and the resulting development of biochemical "-Seq" experiments such as ChIP-Seq, DNase-Seq, and Methyl-Seq over the past decade has given rise to a wealth of predicted enhancers and other cis-regulatory regions (CRMs). These new assays provide a new opportunity to compare the number, location, and possible nature of CRMs that are predicted by various new biochemical techniques to instances of known CRMs, which until recently have primarily been located—for reasons of technological limitations—at a few tens of highly expressed, mostly developmentally-specific genes and the several kilobases (kb) upstream of their promoters. For example, an early surprise in the first ChIP-Seq experiments was that the number of predicted tissue-specific transcription factor-occupied sites outnumbered the number of tissue-specific genes by at least a factor of 10, and that many of these occupied sites were nowhere near developmentally relevant genes. In this thesis, I use the ChIA-PET technique, which preserves factor-containing physical interactions between loci in the genome that are far from each other (10kb-2Mb), where the factors used in this thesis are RNA Polymerase II (pol2) to capture active genes, and separately the developmental transcription factor Myogenin to additionally capture CRMs not at promoters. Overall, I report that (1) the closer together two occupied regions are, the more likely they are to be connected, and (2) that a gene’s activity level is highly correlated with its likelihood of being physically engaged with a distant occupied locus. These lead to the discoveries that occupied regions tend to engage with the active genes nearest to them regardless of the developmental profile of the genes, that many genes engage with multiple individual loci, and that many occupied regions interact with multiple genes, including genes that are not at all related in terms of their expression patterns. Individual elements that have multiple connections likely represent sequential rather than simultaneous interactions, and developmental genes may require more engaged enhancers than genes that are expressed in all cell types. Most excitingly, it is possible that many genes with unchanging expression patterns, including so-called "housekeeping genes," use CRMs; very few such genes have ever been assayed with respect to gene regulation, and they are the vast majority of genes in the genome.

Thesis Files