Causal logic in Boolean modeling of biological networks
Open Access
- Author:
- Maheshwari, Parul
- Graduate Program:
- Physics
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- September 11, 2020
- Committee Members:
- Reka Z Albert, Dissertation Advisor/Co-Advisor
Reka Z Albert, Committee Chair/Co-Chair
Sarah Mary Assmann, Committee Member
Dezhe Jin, Committee Member
Carina Pamela Curto, Outside Member
Richard Wallace Robinett, Program Head/Chair - Keywords:
- Boolean networks
Biological networks
Regulatory functions
Signal transduction network
Attractor
Stable Motifs
Network model
Causal logic
Network motifs
Stomata
guard cells
phosphatases
biomodels
short-term memory
reversibility
feedback loops
oscillating attractor
inference
EMT transition
cancer network
plant biology
calcium signaling - Abstract:
- Cells are complex systems with many biomolecules that interact with each other in multiple ways. These complex interactions lead to intracellular dynamics which in turn govern cellular behaviors, like unique responses to external stimuli. These complex interactions are best represented as intracellular networks, for example, gene regulatory or signal transduction networks. An often-used dynamic modeling framework for these networks, Boolean modeling, can describe their attractors (which correspond to cell types and behaviors) and their trajectories from an initial state (e.g. a resting state) to the attractors, for example in response to an external signal. The existing methods of analyzing Boolean models however do not directly address the causal relationships between distant nodes in the network. In this thesis, I first propose a logic framework, based on categorizing causal relationships as sufficient or necessary, as complementary to the existing analysis of Boolean networks. I identify and explore the properties of complex subnetworks that are distillable into a single logic relationship. I also illustrate how we can use the causal logic analysis method to identify the logic backbone of a biological network, consisting of external signals, self-sustaining cyclic subnetworks (stable motifs), and output nodes. I use the logic framework to identify crucial nodes whose override can drive the system from one attractor to another. Most of this methodology is detailed in Chapter 2 with application to biological networks: the epithelial-to-mesenchymal transition network corresponding to a developmental process exploited in tumor invasion, and the network of abscisic acid induced stomatal closure in plants. While the logic representation identifies the causal relationships between distant nodes and subnetworks, I show in Chapter 3 how we can use the knowledge of causal relationships between biomolecular components to reverse engineer the Boolean network. I present a streamlined network inference method that combines the discovery of the network structure and the identification of the Boolean rules that define the dynamics of the system. I show how we can use the causal logic framework to infer indirect relationships that may not yet have been documented experimentally. I apply this inference method to a well-studied process of hormone signaling in plants, the signaling network underlying abscisic acid (ABA) induced stomatal closure. This signaling network regulates the swelling or shrinking of the guard cell pair surrounding the stomatal pore. Prior Boolean network modeling of this complex cellular process has recapitulated a vast majority of experimental observations regarding the behavior of the guard cell in response to different signals and perturbations. Using the causal logic inference method on this network and comparing it with a carefully curated published model, I show that the inferred network is in good agreement with the published one and it significantly reduces the manual work typically required for network construction. Hence, I conclude that this is an effective method for the inference of biological networks and can help improve our understanding of biological systems by guiding experiments and computations. In Chapters 4 and 5, I further explore and improve the Boolean model of the ABA signal transduction network by using a combination of causal logic analysis, stable motif analysis, and Boolean network simulations. In Chapter 4, I describe a dynamics-preserving network reduction and the identification of predictions that improve the agreement between model and experiment. I show that the causal logic-based reduction method can be used to reduce the network from 81 nodes 153 edges to a condensed 49 nodes and 113 edges while reproducing the predictions of the original model. I then show the utilization of the reduced network to explore cases in which experimental activation of internal nodes in the absence of ABA elicited stomatal closure in wet bench experiments, but not in our in-silico model. The simulations revealed that the addition of a single edge, which allows indirect inhibition of any one of three PP2C protein phosphatases (ABI2, PP2CA, HAB1) by cytosolic Ca2+ elevation, resolves a majority of the discrepancies. This is also supported by the experiments done by our collaborators. These results illustrate how the iteration between model and experiment can improve predictions of highly complex cellular dynamics. In Chapter 5, I show a further improvement of the ABA signal transduction network model such that it captures the biologically expected response of the guard cell in the absence or following the removal of a closure-inducing signal such as ABA or external Ca2+. The expectation from the biological system is reversibility, i.e., the stomata should reopen after the closing signal is removed. The motif succession diagrams help us find that the model’s reversibility is obstructed by the previously assumed persistent activity of four nodes. I find that by introducing time-dependent Boolean functions for these nodes, the model recapitulates stomatal reopening following the removal of a signal. The previous version of the model predicts ∼20% closure in the absence of any signal due to uncertainty regarding the initial conditions of multiple network nodes. I systematically test and adjust these initial conditions to find the minimally restrictive combinations that appropriately result in open stomata in the absence of a closure signal. These results are supported by an analysis of the successive stabilization of feedback motifs in the network, illuminating the system’s dynamic progression towards the open or closed stomata state. This analysis particularly highlights the role of cytosolic calcium oscillations in causing and maintaining stomatal closure. Overall, I illustrate the strength of the Boolean network modeling framework to efficiently capture cellular phenotypes as emergent outcomes of intracellular biological processes.