Understanding Systems-Level Oscillations: Comparative and Network Analysis of Dynamic Phenotypes

Many important biological processes are temporally regulated. For periodic biological processes—such as the cell cycle, the circadian rhythm, and the malaria developmental cycle—temporal ordering of dynamic cellular events is controlled by large programs of periodic gene expression. A large portion of the genome oscillates in these dynamic biological processes (between 20 and 90 percent of the genome for the dynamic biological processes above). These coordinated programs of gene expression are controlled by gene regulatory networks (GRNs) consisting of transcription factors (TFs), kinases, and ubiquitin ligases. GRNs serve to generate and transmit a pulse of transcription, order temporal events, and maintain oscillations over multiple cycles. Historically, uncovering these regulatory mechanisms has taken coordinated effort over decades. An important challenge in biology today is accelerating the rate of discovery for these high-dimensional and complex biological phenomena. To this end, high-dimensional data and complex analytical tools are required. Time-series transcriptomic analyses have uncovered many important insights into dynamic processes, as they enable the characterization of gene-expression profiles for thousands of genes simultaneously. Furthermore, these time-series transcriptomic datasets can be used to infer GRN models from the data. However, these analyses can be complex. Many new computational tools have been developed to enable complex analyses. In Chapters 2, 3, and 4 of this dissertation, I describe methods for analysis of time-series transcriptomic data, network inference, and comparison across experiments. These computational tools were then applied to the Saccharomyces cerevisiae cell cycle to understand the regulation of the cell-cycle period in unfavorable growth conditions. The cell cycle is a vitally important dynamic biological process, which relies on multiple layers of regulation to ensure correct temporal ordering of cell-cycle events and the cell-cycle transcriptional program. Checkpoints monitor cell-cycle progression to ensure correct temporal ordering without catastrophic errors. This ordering is vital to guarantee faithful duplication of the cell. Cell-cycle progression is also affected by environmental conditions. In response to acute environmental stress, the S. cerevisiae cell cycle halts or pauses as a stress responsive and preparative mechanism. In response to chronic environmental stress, the cell-cycle period slows, providing mild stress-resistance. However, the mechanism underlying cell-cycle period control remains under debate. Early studies suggested that this regulation occurs in late G1 via a size or resource threshold at START. More recent studies show that cell-cycle period regulation can occur outside of G1 as well, indicating a need for other regulatory models. One such alternative model comes from the GRN models described above. Turning to other biological oscillators, the circadian period is controlled by the complex circadian GRN. The circadian period is tightly regulated to ensure a 24-hour cycle, matching the natural day-night cycle. However, the circadian period is altered upon experimental perturbation of GRN components, indicating that the circadian GRN controls the period of oscillation. Similarly, perturbation of the cell-cycle GRN alters the cell-cycle period, indicating that the cell-cycle period could similarly be controlled by its regulatory network. Using the computational tools outlined in Chapter 2, 3, and 4, Chapter 5 of this dissertation seeks to understand the mechanisms by which the cell-cycle period slows in response to unfavorable growth conditions. I propose a network model for cell-cycle period control, consisting of stress-regulatory interactions with the cell-cycle GRN. These regulatory models serve as experimental guidance, thus enabling more rapid identification of regulatory mechanisms for complex and high-dimensional biological processes.