Chromatin polymer simulations and bioinformatics-based tools to understand the 3D structure of the genome at allele-specific resolution.

  • Stephen Richer

Student thesis: Doctoral ThesisPhD

Abstract

The human genome spans two metres in length and is packed into a nucleus of no more than a few microns across. It is now widely accepted that the 3D conformation of our DNA is a critical functional property of the genome, influencing a wide range of cellular processes from transcriptional regulation to cell division. In parallel with advances in microscopy, the ongoing development of 3C techniques has enabled researchers to probe the 3D genome with increasing resolution. In particular, the development of HiC has facilitated unbiased genome-wide interrogation of chromatin structure. The novel and unique problems posed by chromatin conformation research have inspired a cross-disciplinary effort employing experimental research, bioinformatics, molecular dynamics and topology, to name a few. Novel computational methods are under continuous development to make sense of these methods' large and complex datasets. Despite this, our understanding of chromatin organisation remains in its relative infancy.

This work describes the development of two distinct computational workflows for analysing experimental HiC data and performing molecular dynamics simulations and in silico HiC. These workflows integrate established methodologies and custom approaches with a robust, modular framework to encourage ease of use and future development. The first workflow, HiCFlow, represents a comprehensive, end-to-end HiC analysis tool uniquely capable of performing de novo allele-specific HiC. The workflow is validated on three key, publicly available human datasets and reveals new insights into genome-wide allele-specific chromatin conformation. The second workflow, HiCSim, utilises coarse-grain polymer simulations to interrogate and visualise simulated chromatin conformation. HiCSim was validated against human HiC data and could reproduce fundamental properties of chromatin conformation with reasonable accuracy. In addition, it is among the first of such approaches to be developed as a user-friendly, configurable workflow that can be employed by non-expert users such as pure experimentalists.

A prevailing philosophy of this work is that computational methods can be powerful, easy to use, and accessible to a wide range of researchers, irrespective of their particular specialism. In enabling non-expert users to directly apply cutting-edge bioinformatics techniques and molecular dynamics simulations to their work, the true benefits of cross-disciplinary research can be realised.
Date of Award28 Jun 2023
Original languageEnglish
Awarding Institution
  • University of Bath
SupervisorAdele Murrell (Supervisor), Laurence Hurst (Supervisor) & Dorothy Buck (Supervisor)

Cite this

'