UCLA

De novo mutations in chromatin modifier genes can lead to a variety of complex developmental syndromes that can have severe consequences for affected patients and their families. In this dissertation we will develop a computational framework for investigating the etiology of this diverse class of disorders, with the underlying motivation being that a deeper and more thorough understanding of the mechanisms underlying these disorders is essential to supporting the development of therapeutics that can improve the quality of life for those affected. In Chapter 1 we will provide background information essential for understanding the work developed in this dissertation. We will begin this chapter with a rather broad overview of the basic biology that grounds our direction of investigation into chromatin modifier syndromes and provide some definitions for key concepts. In Chapter 2 we will then cover in some detail the methods in molecular biology that form the state of the art employed for studying chromatin modifier syndromes. In particular we will look at the various functional genomics assays that are used to measure the transcriptomic and epigenomic effects caused by mutations in chromatin modifier genes. Here we will also give a survey of existing computational methods for the analysis of data generated by these molecular biology assays. In this survey we will highlight several critical gaps that exist in current methods of analysis and note how these hinder investigations into the etiology of chromatin modifier syndromes. This will lead us into the subsequent chapters of the dissertation where we develop methods that address these gaps. In Chapter 3, we will look at the gap that exists in our ability to use existing methods to identify the scale of changes over the genome and develop a method for the analysis of differential DNA methylation that addresses this problem. In Chapter 4 we will look at the limitations of current methods for integrating analysis with the wealth of existing knowledge on the structure of and relationships between biological entities. This limitation we address in our development of a method to weight measures of gene expression specificity based on the similarity structure of the biological entities that compose the underlying sample set. The novel methods that we develop in Chapter 3 and Chapter 4 provide a framework for building a more systems level understanding of the molecular pathology of chromatin modifier syndromes that we believe will be essential in the pursuit of effective treatments and therapies for these diverse and complex disorders. To conclude in Chapter 5, we will summarize our main results and take a brief prospective look at the direction of the field of research into chromatin modifier syndromes making note of promising directions for future research to expand on the work developed here.