Variational Bayesian Methods for Discovering Gene Expression Mechanisms from Single-Cell Transcriptomic Data

PASSCODE: GOODLUCK

Understanding gene expression is vital for fields like cancer research, neuroscience, and medicine. Advances such as single-cell RNA sequencing and spatial transcriptomics enable quantitative analysis of gene expression and cell phenotypes. While current methods can distinguish cell identities based on gene expression, deeper insights into the origins of these differences remain elusive. This dissertation seeks to bridge this gap by addressing three key challenges unique to this domain and summarizing findings from three distinct yet interconnected studies. The first study presents VeloVAE, a variational Bayesian method for recovering temporal information on RNA transcription, splicing, and degradation from single-cell RNA sequencing data. The second study introduces TopoVelo, a graph learning method for uncovering the spatial dynamics of cell migration and differentiation using spatial transcriptomic data. The third study presents ABCDEFG, a differentiable Bayesian causal discovery method designed to reveal causal relationships from single-cell perturbation data. These methods integrate deep learning techniques with domain knowledge to extract hidden information from existing biological data. Through a series of studies, we demonstrate that these approaches achieve state-of-the-art performance and effectively capture biological insights related to the temporal, spatial, and causal mechanisms of gene expression.

CHAIR: Professor Joshua Welch