Maizels, Rory J; (2024) Modelling the dynamics of developmental gene expression with time-resolved single-cell transcriptomics. Doctoral thesis (Ph.D), UCL (University College London).

Text Final Thesis - Rory Maizels.pdf - Other Access restricted to UCL open access staff until 1 July 2025. Download (75MB)

Abstract

Developmental cell fate decisions are driven by the dynamics of gene regulatory net- works. While single-cell genomics can provide highly resolved information on devel- opmental gene expression, this information is only a static snapshot with no details on dynamics. Metabolic labelling and splicing can provide time-resolved information, but current methods have limitations. This study presents experimental and computa- tional methods that overcome these limitations to allow dynamical modelling of gene expression from single-cell data. First, I developed sci-FATE2, an optimised metabolic labelling method that substantially increases data quality. This method was used to profile approximately 45,000 embryonic stem cells differentiating into multiple neural tube identities. To recover dynamics, I developed a variational autoencoder frame- work for velocity inference that outperforms current velocity tools on multiple quan- titative benchmarks. Observing that instantaneous velocity vector fields poorly pre- dict longer-scale cell trajectories, I extended this modelling framework with a neural stochastic differential equation system which models trajectory distributions in latent space. This framework predicts trajectory distributions that can recapitulate dataset distributions, that conserve known biology and that capture dynamical aspects such as decision boundaries between alternative fates and correlative gene regulatory struc- ture. These methods were used to provide a dynamical description of in vitro neural patterning, detailing a sequential decision making process and fate-specific patterns of developmental signaling. Together, these experimental and computational methods aim to recast single-cell analyses from descriptions of observed data distributions to models of the dynamics that generated them, with the goal of providing a new framework for investigating developmental gene regulation and cell fate decisions.

Download activity - last month

Export as XMLJSONCSV

Download activity - last 12 months

Export as XMLJSONCSV

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item