Leveraging Inter-cellular Heterogeneity and RNA Dynamics: Computational Approaches for Investigating Virus Regulation, Host Factors and Pathway Modulators in Single-Cell Transcriptomics

Identifying the factors that influence the development and outcome of biological processes, such as viral infections or drug treatments, often depends on methods that generate snapshots of cellular states, like RNA sequencing (RNA-seq). To address the lack of temporal resolution in these methods, various approaches are employed. Timeseries experiments, for instance, allow for the observation of a sequence of snapshots over time. Additionally, techniques like RNA velocity aim to infer temporal information by exploiting splicing dynamics in RNA-seq datasets. While these methods can significantly enhance data analysis, capturing true temporal dynamics remains challenging and frequently requires dataset-specific approaches. A recently developed method, scSLAM-seq, combines single-cell RNA sequencing (scRNA-seq) with metabolic labeling of newly synthesized RNA, providing true temporal resolution. This innovative approach is a powerful tool to investigate the dynamics of transcription in a wide range of biological processes. In my thesis, I developed both general use and data set specific tools to refine and exploit temporal information in scRNA-seq, SLAM-seq and scSLAM-seq datasets. In chapter 2, I analyze a scRNA-seq dataset of RSV-infected donor cells grown in an air-liquid interface cell culture system to elucidate the dynamics of RSV infection over time, its regulation and exploitation of the host and identified IRF1 as a promising factor for prophylactic and potentially acute treatment of RSV infections. Subsequently, I focused on the SLAM-seq methodology and described previously unknown effects of 4sU-induced quantification bias across various labeling times, concentrations, and cell lines. To address these issues, we developed computational tools to quantify, rescue, and correct mappability-related and biology-based sources of quantification bias, as detailed in Chapter 3. Finally, in Chapter 4, I introduce and develop the concept of Heterogeneity-seq. This method leverages the temporal information provided by scSLAM-seq and intercellular heterogeneity to identify host cell factors that influence the outcome of biological processes. I demonstrate Heterogeneity-seq's ability to successfully elucidate genes that modulate Dexamethasone signaling, as well as pro- and antiviral factors in the context of MCMV infection.