| dc.description.abstract | Nucleosome positioning is involved in a variety of cellular processes, and it provides a
likely substrate for species evolution and may play roles in human disease. However, many
fundamental aspects of nucleosome positioning remain controversial, such as the relative
importance of underlying sequence features, genomic neighbourhood and trans-acting
factors.
In this thesis, I have focused on analyses of the divergence and conservation of
nucleosome positioning, associated substitution spectra, and the interplay between them. I
have investigated the extent to which nucleosome positioning patterns change following the
duplication of a DNA sequence and its insertion into a new genomic region within the same
species, by assessing the relative nucleosome positioning between paralogous regions in both
the human (using in vitro and in vivo datasets) and yeast (in vivo) genomes. I observed that
the positioning of paralogous nucleosomes is generally well conserved and detected a strong
rotational preference where nucleosome positioning has diverged. I have also found, in all
datasets, that DNA sequence features appear to be more important than local chromosomal
environments in nucleosome positioning evolution, while controlling for trans-acting factors
that can potentially confound inter-species comparisons.
I have also examined the relationships between chromatin structure and DNA sequence
variation, with a particular focus on the spectra of (germline and somatic) substitutions seen
in human diseases. Both somatic and germline substitutions are found to be enriched at
sequences coinciding with nucleosome cores. In addition, transitions appear to be enriched in
germline relative to somatic substitutions at nucleosome core regions. This difference in
transition to transversion ratio is also seen at transcription start sites (TSSs) genome wide.
However, the contrasts seen between somatic and germline mutational spectra do not appear
to be attributable to alterations in nucleosome positioning between cell types. Examination of
multiple human nucleosome positioning datasets shows conserved positioning across TSSs
and strongly conserved global phasing between 4 cancer cell lines and 7 non-cancer cell
lines. This suggests that the particular mutational profiles seen for somatic and germline cells
occur upon a common landscape of conserved chromatin structure.
I extended my studies of mutational spectra by analysing genome sequencing data from
various tissues in a cohort of individuals to identify human somatic mutations. This allowed
an assessment of the relationship between age and mutation accumulation and a search for
inherited genetic variants linked to high somatic mutation rates. A list of candidate germline
variants that potentially predispose to increased somatic mutation rates was the outcome.
Together these analyses contribute to an integrated view of genome evolution,
encompassing the divergence of DNA sequence and chromatin structure, and explorations of
how they may interact in human disease. | en |
