PDF, English - main document Restricted access: Repository staff only until 27 June 2025. Login+Download (10MB) | Lizenz: Lineage dynamics of haematopoiesis: insights from single-cell multi-omics and lineage tracing von Greco, Alessandro steht unter einer Creative Commons Attribution-NoDerivatives 3.0 Germany

Abstract

In haematopoiesis research, the description of the molecular states involved in differentiation from tip stem cells to unipotent progenitors is an ongoing effort that incorporates newly available technologies.

In particular, the application of single-cell sequencing technologies enables the dissection of transcriptional, epigenetic and immunophenotypic heterogeneity of stem and progenitor cells. As importantly, in-vivo fate mapping allows to trace the progeny of stem cells and their progeny and infer dynamical properties of the haematopoietic system.

However, a unified view of haematopoietic states that incorporates molecular heterogeneity in the context of differentiation relationships between subsets is still lacking.

In this thesis, I integrate information from multiple assays to obtain a high-resolution description of haematopoietic heterogeneity and link it to differentiation insights from lineage tracing evidence using statistical testing. The results presented here support a hierarchical model of the haematopoietic system, in which tip stem cells undergo an ordered sequence of regulatory events involving exit from a primitive state, proliferative activation, followed by intermediate lineage restriction steps that result in uni-potent progenitors.

In Chapter 2, fate mapping and index-sorted scRNA-seq data are combined to investigate how transcriptional and immuno-phenotypical heterogeneity affect the direct differentiation pathway between LT-HSCs and MkPs.

In the following chapters, a multi-omic dataset featuring paired scRNA-seq and scATAC-seq layers is analysed to investigate the mappability between the two and unveil modality-specific states related to proliferation, extrinsic signalling, and three-dimensional chromatin re-modelling.

Next, I utilise probabilistic trajectory inference and validate it using transcriptional mapping to lineage tracing datasets. The output of this procedure is used to inform a lineage potential model that enables statistical testing of fate association and thus the degree of hierarchy in lineage choices. Unbiased grouping of mature cell fates results in a branching model that features early lineage split within the stem cell compartment into intermediate oligo-potent progenitors before further specification into uni-potent progenitors.

To gain a molecular understanding of the detected branching events, I make systematic use of differential expression and differential accessibility analysis to investigate the overall mechanisms that govern commitment at both transcriptional and chromatin level, thus use mathematical modelling to detect the hierarchical order of regulatory events. Finally, I utilise a recently developed gene regulatory network inference algorithm to reveal a highly dynamical regulatory landscape that links haematopoietic specification to key transcription factors.

In summary, the work presented in this thesis makes use of bioinformatic analysis, mathematical modelling and lineage tracing to obtain a robust description of the cellular states that comprise the haematopoietic system and their relation to differentiation potential and proposes a computational framework that can be used to enable the quantitative description of differentiation systems starting from single-cell sequencing data.