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.
Monday, October 28, 2019
SysBio19 Asgn_8_Due_Tuesday_43767__Liu_article_in_Control_Folder_2019_10_29 Lead: Javier
Network Observability: Read carefully Liu article_in_Control_Folder Liu, Y. Y., et al. (2013). "Observability of complex systems." PNAS 110(7): 2460-2465.. 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 Jaview to explain in class? 4) How does this relate to your project?
Subscribe to:
Post Comments (Atom)
1) What did you already know?
ReplyDeleteI had an understanding of the complex nature of networks and observing their state based on limited data
2) What did you learn?
I learned about how to deconstruct a network/inference diagram into the minimum number of "sensor" nodes needed to determine the network's state by identifying SCCs
3) What is your most pressing question that you would like Javier to explain in class?
-How accurate does your proposed network have to be-- will one inaccurate relationship throw you off?
-I do not fully understand how you can obtain complete observability without knowing the kinetic constants
-How does noise affect the system
4) How does this relate to your project?
If we create (or find) a network that involves the main pathways related to dedifferentiation, we can identify the minimum number of sensor nodes needed to achieve complete observability of the system, which could potentially be used to help identify what state a cell is in and if it is headed in the right direction.
1) What did you already know?
ReplyDeleteI knew that systems are often difficult to observe internally. I also knew what was meant by the terms "necessary" and "sufficient."
2) What did you learn?
I learned that the internal state of a system can be reconstructed if the minimum number of strongly connected components are monitored. This is possible through the graphical approach (GA), which reduces the observability problem to a property of the diagram drawn.
3) What is your most pressing question that you would like Javier to explain in class?
Can you give succinct definitions and differences between the terms "SCC," "root SCC," and "sensor node"?
4) How does this relate to your project?
When trying to construct a Boolean network to describe a gene regulatory network, this work tells us that we only need to monitor the nodes of the gene regulatory network that are necessary and sufficient to describe the entire internal state of the system. Therefore, we can greatly reduce the computational load required to construct the system and the experimental load to verify the system.
0) Knew: graphical approach to viewing networks and breaking them down into strongly connected components (SCCs)
ReplyDelete1) Learned: There is a mathematical definition for observability and a way to determine which nodes in a network must be monitored for full observability.
2) Question: What if the identified nodes are not measurable in your desired system? i.e. if you want to only measure metabolites but a non-secreted protein/gene is identified as a critical node for full observability. Is less than complete observability acceptable?
3) Presentation: examples of choosing nodes to monitor in a system modeled by a network of interacting components. Is complete observability achieved by Liu's standards?
4) Thoughts: I think we visited this paper last year, but the concept hits home a little more this time, especially in thinking about optimal control of systems where your measurable outputs are limited.
1) necessary and sufficient criteria for fully understanding a system and how to potentially observe it with minimal data
ReplyDelete2) how to decompose a interference/network diagram into sensor nodes, mathematical definition of observability
3) Simple definitions and biological examples of some of the terms, a simple example of decomposing a network diagram
4) how to reduce computational load when trying to understand mechanism of disease/differentiation
1) What did you already know?
ReplyDeleteThat biochemical reaction pathways and networks could be mapped to explore relationships between reactions and correlate those to system function
2) What did you learn?
That sensors do not have to be placed at every node throughout a biochemical reaction map pathway, but the reaction can be monitored looking at only a few key nodes.
3) What is your most pressing question that you would like Javier to explain in class?
Could you go into more detail and give examples of how root versus non-root strongly connected components are distinct?
4) How does this relate to your project?
Rather than waste effort in observing all reactions taking place in a cell differentiation network, we should instead focus on monitoring a few and extrapolating those results to understand differentiation and biological behavior with less effort.
1.) What did you already know?
ReplyDeleteMy background in control is mostly limited to the 'controllability' & 'Leduc' papers. I was aware that most variables in complex systems as the ones we find in biology are rarely independent.
2.) What did you learn?
There is a minimum set of variables that need to be monitored in order to 'observe' the network.
3.) What is your most pressing question that you would like Javier to explain in class?
How easy is it to implement a technique like this having little to know prior experience.
4.) How does this relate to you project?
This could solve the issue of requiring too many measurements in order to observe the state of a system, aka the reprogramming of our cells. Thus we could possibly implement a targeted sequencing/mass spec with higher accuracy and maybe more affordable.
1) What did you already know?
ReplyDeleteBasics of observability and graph terminology.
2) What did you learn?
SCCs can be used for observing biochemical networks and there is a lack of symmetry in biochemical networks.
3) What is your most pressing question that you would like Javier to explain in class?
I am confused by how the graph in Fig 1c includes links such as A and B when A doesn't produce B or vice versa. Why are these links included?
4) How does this relate to your project?
Oberservability will be important for controlling stem cell differentiation.
1) What did you already know?
ReplyDeleteIt is both necessary and difficult if not impossible to understand the internal state of a system to fully predict the outcomes.
2) What did you learn?
Something is observable if we can reconstruct the system’s complete internal state from its outputs. This can be done by identifying the minimum set of sensors where we can use those measurements to determine all other variables.
3) What is your most pressing question that you would like Jaview to explain in class?
In the situation where you have N species involved in R reactions, where don't you need to do N*R perturbations while using this method? How can you tell which variables are redundant?
4) How does this relate to your project?
Many of the SCCs for iPSC production have already been experimentally determined and there's lots of data as well, so this could be a way to avoid doing more experiments to determine best efficacy or a more "pure" stem state.
1) I did not know much of the material in the paper. I knew what observability is, but not much else.
ReplyDelete2) The presence of SCCs and how you can observe a system through a couple sensor nodes rather than the entire network.
3) Why are sensor nodes interchangeable in the SCCs?
4) We could model GRNs as a boolean network with sensor nodes or driver nodes instead of including all genes in the network.
1) What did you already know?
ReplyDeleteI knew the general principle behind observability of a system (reconstruct internal state from outputs) and the inherent limitations due to complexity and sensors in order to observe a system.
2) What did you learn?
There is a minimum set of sensors to observe a system. Not every piece needs to be monitored and we can construct our observation through analysis with sensors not always on every node (remove the redundancy)
3) What is your most pressing question that you would like Javier to explain in class?
How do we select which nodes to monitor/which ones are redundant?
4) How does this relate to your project?
This paper in conjunction with the other Liu paper can be useful. We are trying to control and then observe a system. To do so effectively, these papers propose a guideline for accomplishing it.