Single cell mitochondrial dna: bioinformatics and statistical approaches
File(s)
Author(s)
Marshall, Aidan
Type
Thesis
Abstract
New technologies drive scientific discoveries. Recently, genomics sequencing has rapidly advanced, opening numerous doors for understanding human health and disease. Sequencing techniques are capable of producing prodigious amounts of raw sequencing data routinely, and their output is only increasing. Developments in computational and statistical methods for processing and analyzing this data have striven to keep pace with this data growth. This thesis builds on these developments in the context of studying mitochondrial DNA (mtDNA) and understanding the role it plays in health and disease made possible through nascent single-cell genomics technologies.
We begin by detailing the landscape of single-cell sequencing techniques and existing work applying these techniques to study cellular mtDNA heterogeneity. Multitudes of single-cell datasets are publicly available and current analyses largely ignore the abundant mitochondrial content of their data. Importantly, single-cell datasets allow us to gain an understanding of the role mtDNA plays that is missing in bulk genomic datasets. Previous research has linked mtDNA mutations to ageing pathology. Using the Moran model, we show that post-mitotic cells should accumulate mtDNA mutations, reaching a pathological threshold on the timescale of a human lifespan. Such mutational burden is only seen at the single-cell level, which we term cryptic as it is unobservable in bulk datasets. A pipeline we developed leveraged available singlecell data, confirming that empirical results corroborate theoretical expectations. Moreover, we link the presence of cryptic mutations to many classic hallmarks of ageing.
We then use novel scATAC data to explore the landscape of widespread mtDNA mutations in the developing brain. Such samples are frozen, limiting the mtDNA content that can be captured from each cell. To overcome this, we take advantage of tools from Bayesian inference. This approach allows us to acquire compelling functional insight from low-depth data, whilst faithfully representing the inherent uncertainty.
We begin by detailing the landscape of single-cell sequencing techniques and existing work applying these techniques to study cellular mtDNA heterogeneity. Multitudes of single-cell datasets are publicly available and current analyses largely ignore the abundant mitochondrial content of their data. Importantly, single-cell datasets allow us to gain an understanding of the role mtDNA plays that is missing in bulk genomic datasets. Previous research has linked mtDNA mutations to ageing pathology. Using the Moran model, we show that post-mitotic cells should accumulate mtDNA mutations, reaching a pathological threshold on the timescale of a human lifespan. Such mutational burden is only seen at the single-cell level, which we term cryptic as it is unobservable in bulk datasets. A pipeline we developed leveraged available singlecell data, confirming that empirical results corroborate theoretical expectations. Moreover, we link the presence of cryptic mutations to many classic hallmarks of ageing.
We then use novel scATAC data to explore the landscape of widespread mtDNA mutations in the developing brain. Such samples are frozen, limiting the mtDNA content that can be captured from each cell. To overcome this, we take advantage of tools from Bayesian inference. This approach allows us to acquire compelling functional insight from low-depth data, whilst faithfully representing the inherent uncertainty.
Version
Open Access
Date Issued
2024-03-27
Date Awarded
2025-04-01
Copyright Statement
Attribution-NonCommercial 4.0 International Licence (CC BY-NC)
License URL
Advisor
Jones, Nick
Publisher Department
Department of Mathematics
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)