Mining of Next Generation Sequencing (NGS) and Affinity-Purification Mass Spectrometry (AP-MS) data to identify chromatin targets of epigenetic regulators
File(s)
Author(s)
Lukauskas, Saulius
Type
Thesis
Abstract
In eukaryotes, modifications of DNA molecules and core histone proteins regulate core nuclear functions such as DNA transcription, DNA replication, and DNA repair. Individual modifications are associated with regions of active and repressed transcription, and their readers, writers, and erasers are known. In combination, these modifications are assumed to form a ‘code’ describing the chromatin regulation. The specifics of this combinatorial ‘code’, however, are still unknown. Recent advances in technology have allowed this regulation by nucleosomal modifications to be studied experimentally in high throughput. Such studies, however, face an information overload as more data is being produced than can be analysed. In this thesis, we attempt to provide methods to study this large-scale data globally. To do this, we describe a number of data mining approaches to extract, summarise and link the data in order to elucidate the language of nucleosomal modifications.
We base our analyses on a novel SILAC nucleosome affinity purification (SNAP) dataset produced in our laboratory, as well as data from Next-Generation Sequencing (NGS) studies. We apply a modified CLR network inference algorithm to predict novel interactions between epigenetic readers. We apply Factor Analysis (FA) to describe the key constituents of the code that drive the epigenetic reader binding. We further show that linear combination of individual responses to modifications can describe a significant proportion of the reader behaviour in vitro, and use NGS datasets to verify them in vivo. We find discrepancies between the in vitro and in vivo results and use Bernoulli Mixture Models (BMM) to analyse them. This provides us with a direction to improve the analysis and experimental methods.
Together, the results presented in this thesis provide a global overview of major aspects of the regulation by nucleosomal modifications, which brings us a step forward towards cracking the histone code.
We base our analyses on a novel SILAC nucleosome affinity purification (SNAP) dataset produced in our laboratory, as well as data from Next-Generation Sequencing (NGS) studies. We apply a modified CLR network inference algorithm to predict novel interactions between epigenetic readers. We apply Factor Analysis (FA) to describe the key constituents of the code that drive the epigenetic reader binding. We further show that linear combination of individual responses to modifications can describe a significant proportion of the reader behaviour in vitro, and use NGS datasets to verify them in vivo. We find discrepancies between the in vitro and in vivo results and use Bernoulli Mixture Models (BMM) to analyse them. This provides us with a direction to improve the analysis and experimental methods.
Together, the results presented in this thesis provide a global overview of major aspects of the regulation by nucleosomal modifications, which brings us a step forward towards cracking the histone code.
Version
Open Access
Date Issued
2018-10
Date Awarded
2019-05
Copyright Statement
Creative Commons Attribution NonCommercial Licence
Advisor
DiMaggio, Peter
Sponsor
Biotechnology and Biological Sciences Research Council (Great Britain)
Grant Number
1377205
Publisher Department
Chemical Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)