Haematopoietic stem cells are responsible for producing and sustaining the diverse array of cell types present in the adult blood system. This complex process requires the strict regulation of haematopoietic fate decisions and differentiation trajectories in order to maintain a healthy state. Haematological malignancies such as leukaemia are associated with various perturbations that disrupt this regulation and drive aberrant cell fate decisions, leading to disease. Much of this dysregulation is proposed to occur at the transcriptional level, and recent technological advancements in single-cell sequencing have made it possible to study the transcriptional effects of leukaemic perturbations at the scale of individual haematopoietic stem and progenitor cells. However, the mechanisms through which specific perturbations lead to dysregulation of the blood system remain poorly understood.

The primary aim of this work was to build an integrative computational framework for the analysis and comparison of leukaemic perturbations of the murine blood system as measured by single-cell RNA sequencing. Presented in Chapter 3, this framework aims to dissect the perturbation response across different scales – from individual genes to specific progenitor cell types to the entire blood system – and allow informative comparisons to be made about the similarities and differences between several perturbations. In total, eight genetic perturbations known to associate with leukaemia were analysed, resulting in novel biological insights concerning the behaviour of coordinated gene modules and the cellular abundance shifts driven by them.

As many leukaemic drivers act directly upon the most immature long-term haematopoietic stem cells, a highly targeted analysis of these cells was performed across the leukaemic perturbations. In Chapter 4 a novel computational pipeline was built to link FACS-sorted cell populations and single-cell transcriptional landscapes. Using this, the cellular and molecular responses of the perturbations were investigated, resulting in several novel hypotheses. For example, the data suggests that many leukaemic perturbations gain a competitive advantage against wild-type cells by pushing their MPP1 cells into more active states. Additionally the data suggests that increases in the transcriptional variability of blood stem cells is associated with pro-erythroid fate decision shifts and vice-versa.

Many different types of haematopoietic perturbations exist and can drive disease progression in the blood system. Chapter 5 focuses on single-cell RNA sequencing data from three further perturbations in various settings, including an infection model of Malaria and a model susceptible to endogenous DNA damage by aldehydes. These analyses have driven and validated bodies of experimental work, and comparing them to the previously described perturbation models highlighted both conserved changes and differences in the response of the haematopoietic system across different perturbation settings.

The final project aimed to improve upon current computational methods for cellular trajectory inference from single-cell data. Whilst high-throughput experiments allow for the sequencing of large cell numbers, this is balanced by the sparse and noisy nature of the returned data. Current methods perform poorly on such datasets and either cannot deal with large cell numbers or cannot extract enough relevant signal from sparse count matrices. A new computational tool was designed to work best on these large, sparse datasets, and infer the most likely cellular trajectories through snapshot sequencing data using an iterative process. In Chapter 6 this algorithm was applied to different systems including adult haematopoiesis, and was compared to state-of-the-art methods.

Overall, this thesis has investigated the transcriptional consequences of numerous preleukaemic perturbations on the haematopoietic stem and progenitor cell compartment at the single-cell level. New methods have been built for integration of single-cell perturbation experiments and their analysis across different biological scales. This has revealed novel biological insights regarding the mechanisms underpinning leukaemic transformation of the blood system.