Identifying Dynamic Genes and Regulatory Elements Using Computational Analyses of Multi-omics Data

Gene expression is meticulously coordinated in distinct cell types, requiring precise temporal and spatial regulation. Multi-omics analyses allow for the unraveling of the complex landscape of gene regulation within biological systems. It provides a comprehensive vantage point from which we can comprehend how genes are controlled and function within specific biological contexts. In this thesis, three distinct projects have been developed, all centered around investigating dynamic gene expression through computational analyses of multi-omics data. In the first project, the computational platform TrendCatcher was developed to identify dynamic differentially expressed genes (DDEGs) in time-course bulk RNA-seq and single-cell RNA-seq data. Through this tool, we successfully unveiled unique dynamic transcriptional gene signatures and associated biological processes within longitudinal datasets. Notably, it shed light on the early and persistent activation of neutrophils and coagulation pathways as a hallmark of patients progressing to severe COVID-19. The second project focused on the exploration of dynamic gene expression programs and regulatory elements, specifically transcription factors, during hepatocyte differentiation from pluripotent stem cells. Leveraging time-course ATAC-seq data and RNA-seq data integration, we identified candidate transcription factors that drive gene expression programs at each differentiation stage. Furthermore, using TrendCatcher, we inferred the dynamic trajectories of transcription factor activation. In the third project, we developed ProxSeek, a novel computational platform designed to harness a promoter-centered interactome constructed from Hi-C data to generate a hierarchy of cis-eQTLs. This innovative approach showcased that the promoter-centered interactome serves as a hotspot for cis-eQTLs, and ProxSeek effectively prioritizes these cis-eQTLs (prox-eQTLs), which are enriched in regulatory regions and exhibit higher effect sizes. ProxSeek has the potential to complement existing cis-eQTL studies, aiding in the prioritization of regulatory genetic variants and deepening our understanding of the intricate interplay between genetics and gene expression. Collectively, these projects underscore the value of multi-omics approaches in exploring gene expression regulation, particularly within the framework of temporal processes.