The Systems Biology of COVID-19 and the SARS-CoV-2 virus. The class will build a foundation that includes the emergence of complexity, simple biological subsystems, their reductionist and equivalent toy and organ-chip models, and the measurements required to specify model architecture and parameters. Applications to biology, physiology, medicine, chemical and biological defense, pharmacology, drug discovery, and toxicology. UGrad: PHYS 240 01 and BME 290B; Grad: PHYS 326 and BME 395C.
Thursday, November 7, 2019
SysBio19 Asgn_9_Due_5:00PM Monday Nov 11: Gates article in Boolean_Read folder_2019_11_12 Lead: Kyle
Structural and Dynamic Network Control: Read
carefully Gates article in Boolean_Read folder Gates and Rocha, SciRpts
(2016) "Control of complex networks requires both structure and
dynamics.". Answer the questions: 1) What did you already know? 2) What
did you learn? 3) What is your most pressing question that you would
like Kyle to explain in class? 4) How does this relate to your project?
Subscribe to:
Post Comments (Atom)
1) What did you already know?
ReplyDeleteI knew that methodologies based on the structure of networks have been proposed to determine driver nodes/controllability of networks.
2) What did you learn?
Structural analysis alone is often not sufficient to determine driver nodes/controllability and network dynamical properties need to be considered as well
3) What is your most pressing question that you would like Kyle to explain in class?
Would like more explanation about how you create a CTSG and CAG.
Explanation of the terms introduced in the section titled "Boolean Network Ensembles"- contingent vs noncontingent subsets, canalizing functions, effective structures
4) How does this relate to your project?
Would need to ensure that network matches assumptions of structure-only methods in order to predict control in this way
1) From Liu paper: studying structure of a control system will tell us which nodes are critical/normal/redundant or a driver node
ReplyDelete2) Driver nodes determined by doing a structure analysis might not give a complete picture of controllability or might not be the simplest possible structure
3) If we are studying a new system, would you suggest studying what connections there are (structure) or how different things interact (dynamics) first? What is the methodology of applying what we have learned to a new network?
4) Combine with what we know about testing structures with this paper about testing dynamics to get a complete picture on how we can control a system.
1) What did you already know?
ReplyDeleteI knew the methodologies around structural controllability, but I didn't realize how much the limited nature of boolean networks would limit the ability of the model to predict the number of driver nodes.
2) What did you learn?
I learned about many different types of boolean networks and some more ways to try and control a system. I also learned how just structure based methods for controllability have large margins of error when it comes to trying and predicting controllability in some systems so you need to add dynamic interactions to improve the model.
3) What is your most pressing question that you would like Kyle to explain in class?
Can you explain Controlled States Transition Graph and contingent and non contingent subsets?
4) How does this relate to your project?
This allows a more in depth look into what needs to be done in order to create a boolean network that can actually allow for a better understanding of the system like adding in dynamic interactions. This also introduced many different types of boolean networks that could be used to try and illustrate the states better.
1) What did you already know?
ReplyDeleteUp to this point I'm familiar with network structure from Liu and Boolean from the papers we discussed on the subject.
2) What did you learn?
The primary bit taken away from the article is the importance of dynamics on the overall system's function. More importantly, how dynamics limit the state space that can be reached by any one control scheme i.e. certain states cannot be reached because they are dynamically prevented from doing so.
3) What is your most pressing question you would like Kyle to explain in class?
The most alarming part of the articles is their discussion of driver nodes and the inaccuracy of Liu in certain situations. When is Liu's structure only approach a sufficient approximation? In the case where Liu overestimates, by how much, and what are the criteria for underestimation?
4) How does this relate to your project?
My current interest for our project is related to the function of tau protein signaling networks in brain cells. It's feasible we could apply this approach (or Liu's) to this network to simulate its failure to act normally.
1.) What did you already know?
ReplyDeleteMy main background for this article was the previous paper presented from Liu and how they claimed that network topology alone was enough to determine the critical nodes necessary for controlling a network.
2.) What did you learn?
The importance of canalization in determining the controllability of a boolean network and how they can reduce the size of the effective structural graph, leading to failed predictions from structure based methods.
3.) What is your most pressing question you would like Kyle to explain in class?
I believe I conceptually understood what a canalizing function is and why it is relevant to control, but I still am having trouble understanding it in some of the terms used in the article. Specifically what edges of a network represent and how canalizing functions affect them.
4.) How does this relate to your project?
This article further developed my working understanding of boolean networks and how to control them. With a better understanding of the methods, this could be implemented in our personal control projects.
1) What did you already know?
ReplyDeleteComplex networks can be characterized by their observability and controllability, but there is some debate over whether controllability can be deduced simply from a network's structure, without respect to its dynamics.
2) What did you learn?
Structural controllability (SC) methods both underestimate and overestimate the minimum number of driver nodes required to control a system, depending on the system's dynamics.
3) What is your most pressing question you would like Kyle to explain in class?
Can you define canalization again and talk about its impact on system dynamics, robustness, and controllability?
4) How does this relate to your project?
This article tempers my expectations for the degree of impact of each of our projects. Some papers we have read recently suggested that identifying the structure of a gene regulatory network would be sufficient to estimate controllability of the system. However, Gates suggests that we need to know something about the dynamics of the system, which requires in vitro experimentation. Therefore, we will either have to narrow the scope of our projects or expand our preliminary literature search to gather data on network dynamics.
1) What did you already know?
ReplyDeleteI presented the Lui controllability paper and learned about the dynamics of types of control systems in that. From other papers, I learned about how complex networks can be controlled and how Boolean networks work.
2) What did you learn?
I learned about how the approaches in the Liu paper may not necessarily be enough to control a network. For example, you may need to control more or less nodes in order to reach full controllability. I learned about cannalizing functions how to understand boolean networks with them.
3) What is your most pressing question that you would like Kyle to explain in class?
Does this paper disprove the Kalman rank controllability condition?
4) How does this relate to your project?
Once we generate a boolean GRN for our individual projects, it would be interesting to see which nodes need to be controlled in order to control the whole network. Based on conclusions from the Liu paper, you probably will not want to control the hub nodes which in our network will likely be transcription factors.
1. Knew: Liu argues that network controllability can be estimated by topology; Gates argues that both topology and dynamics are necessary.
ReplyDelete2. Learned: Structural control approaches alone may under/over-estimate the driver nodes of a system
3. Question: How are boolean functions used to test controllability chosen/defined here? What role does canalization play in this?
4. Project: Understanding Gates and Liu's methods for determining controllability is pivotal to evaluating our chosen networks for the semester project(s).
1) What did you already know?
ReplyDeleteConcepts around controlability and nodal analysis of Boolean Networks were familiar to me. I was also aware of how Boolean Network structure could be used to simulate cellular functions and pathways.
2) What did you learn?
That structure predicts dynamics but does not control cell fate and that structural network controlability theory is flawed. The Structure of the Boolean Network does not do well at approximating the realty of cellular complexity.
3) What is your most pressing question that you would like Kyle to explain in class?
I have two general concepts I would like explained better. The first is dynamics. I can visualize a structure - is dynamics the term used to describe what actually occurs, or can we model dynamics apart from structure to create a prediction model that more acurately suites reality. Also, could you explain canalization's role in determining structure - dynamic relationships? The paper kept noting its importance, but I failed to see how it was actually implemented.
4) How does this relate to your project?
When creating a Boolean Network for our project, finding a way to incorporate dynamics accurately into the structure will be a necessary challenge. We Will have to better define what dynamics look like in our network and create a simulation to replicate that.
1) What did you already know?
ReplyDeleteTypes of networks and how they are controlled differently.
2) What did you learn?
Structural control predictions can both underestimate or overestimate the number of driver variables in these systems. These approaches also fail to predict which sets of variables best control dynamics
3) What is your most pressing question that you would like Kyle to explain in class?
What is the difference in the analysis of a network using linear and nonlinear approaches
4) How does this relate to your project?
There seems to be important context for all boolean networks, which we will all be using for our projects
^^ I forgot to press publish before. When I tried to do it later it logged me out and published as unknown but that's me
ReplyDeleteNow it should be showing up as Blake oops
ReplyDelete* I am sorry this is late, I had it written, but I realized that I forgot to submit it into the blog.
ReplyDelete1) What did you already know?
From previous papers that we have read, I understand the basic concepts of the structure of networks and understanding them through different nodes, controlling them through things like driver nodes
2) What did you learn?
The structural prediction are insufficient as they do not accurately estimate the nodes necessary for controlling/observing a network.
3) What is your most pressing question that you would like Kyle to explain in class?
- not sure I completely understand what exactly a canalizing function is.
-How do we select canalization to improve controllability?
4) How does this relate to your project?
When we are looking to develop our own boolean networks, we can look to this paper and the Liu papers to get a complete picture on how to incorporate testing dynamics and controlling a system.