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, August 25, 2016
Subscribe to:
Post Comments (Atom)
0. Knew:
ReplyDeleteI have had a little bit of experience with cellular signaling. The research group I am a part of has been working with a cell and developmental biology group looking at the Calcium signals in response to laser ablation wounds in epithelial tissue of fruit fly pupae. In looking at what type of signal could initiate the release of calcium, a couple of back of the envelope calculations have been made to narrow down the possibilities given the spatial and temporal time scales of the data.
I just recently was made aware of how important these back of the envelope calculations can be in biological systems. In the draft of the book “Cell Biology by the Numbers” by Ron Milo and Rob Phillips recommended for me to check out by Dr. Weil in the intro to chemical and physical biology class, it states about these types of calculations “We often want to do “quick and dirty” analyses to estimate time scales, rates, energy scales or other interesting biological parameters as a sanity check to see if some observation or claim makes sense.” This idea is what drives the first section of this paper.
I was aware of two of the four experimental techniques mentioned on page 106. A student in the cell and developmental biology group is using RNAi in Drosophila melanogaster pupae to determine the functions of various receptors in the calcium signaling. Also, the calcium signals I have been analyzing use the fluorescent protein GCaMP6 which binds to the calcium, allowing us to determine how the concentration of calcium diffuses.
Although the paper did not discuss graph theory, I took a course on it at my undergraduate institution, and panel a in box 2 as well as figure two scream graph theory to me. I am sure it plays a huge role in all of this mathematically.
1. Learned:
The main thing I learned about from this review is the steps in the process, or cycle, of network reconstruction. Although figure 1 is specifically about integrative and iterative process of cellular signaling network reconstruction, I believe it captures the essence of this review. There is so much data out there on so many different signaling networks (or any biological process in general). By starting with a base understanding of the data one can use orders of magnitude estimates along with a desired complexity to determine what path to take mathematically. This then leads to new interpretations of existing data and formulation of new data, which then allows more network reconstruction, and the process repeats itself. This is not a trivial task at all, but it is a nice road map to follow.
Excellent - I've ordered a copy of “Cell Biology by the Numbers.” The inverse network problem requires a lot more than order-of-magnitude estimates.
DeleteThis paper was well organized enough that learning new terms was very simple (definitions of many terms are located in the left margin). So I can confidently say I have learned a lot of new definitions and signaling functions!
ReplyDelete2. Pressing Questions:
Why was graph theory not mentioned in this paper (except in the definition of spectral analysis)? The author specifically states on page 104 in discussing the modular description of signaling networks, “These considerations compel seeking a definition of modules from a mathematical, network-based perspective rather than subjective reasoning” This is what graph theory is all about. I know the paper is not about graph theory, but I got the vibe that the author views analyzing these networks in this way as something that still needs to be greatly developed mathematically, when in fact it already has been.
3. Presentation Topic:
It would be nice to hear more about the mathematical side of network reconstruction. The author does a nice job at citing or mentioning that certain types of mathematical analysis techniques have been used, but I was left wanting more. I want to know more specifically what is being used and how it leads to the various results mentioned in the paper. Of course this presentation would not be about this paper, but rather a supplement to it.
Another presentation topic could be presenting a specific case where network reconstruction has been used and the steps that were taken to achieve it. The author does a good job at giving examples of when certain techniques have been used, but different examples are used each time (as a review should). Being able to see a single example all the way through would be a good way to summarize the subject of this review.
4. Thoughts:
This paper was easier for me to get through than the previous one. This is most likely because I had more prior knowledge than with the previous article. In any case, I enjoyed this article very much, and I am enjoying the format of this class so far. I cannot wait for the discussions!
This comment has been removed by the author.
DeleteWe will be working on this over the course of the semester!
Delete0. Knew
ReplyDeleteThanks to undergrad cell biology, I knew a little bit about cell signaling—mainly that it was very complicated, had an immense amount of variability, and, in the case of some pathways, a fair amount of redundancy within the pathway itself. However, any knowledge I had of specific pathways had long been forgotten.
I have a small amount of experience with several of the data collection techniques mentioned within this paper—chromatin immunoprecipitation (ChIP), although I had previously encountered it in ChIP-sec technology, which is different from the ChIP-chip assays mentioned in this paper. While I have heard of RNAi before, my knowledge of it is far from complete. However, I am very familiar with many types of fluorescence imaging technologies, particularly those used for protein localization, as my previous research was very heavily focused on fluorescence imaging.
I had also encountered similar “order of magnitude” calculations and thus, was aware of their usefulness in estimating the approximate size of large data sets.
1. Learned
While I was aware of the complexities of signaling networks, and would assume that reconstructing them would be equally complex, I had never given much thought to exactly how one would accomplish this task. I learned that due to the vast amount of intracellular components, many of which depend on other intra- and extra- cellular components or events, it is difficult to isolate the mechanisms by which a single signal operates. Because these many components and interactions are so interconnected, it is near impossible to deconstruct a network down to individual reactions, since most reactions and pathways exist due to a prior one. Additionally, if you try to deconstruct too much, an experiment might accidentally result in “cross-talk” between network components that do not interact in the actual signaling network, leading to new reactions that are not part of the original network and thus, an incorrect understanding of that pathway.
These biochemical complexities lead to two different approaches for network reconstruction—integrative and iterative. Beyond that, there are several different ways to approach the subsequent mathematical analysis of networks and their properties.
2. Pressing
I still don’t have a firm grasp of the difference between nodes, modules and pathways (Box 2), how you can use these descriptions for network signaling examples, or why one might be more accurate than another.
3. Presentation
I would like more details about nodular vs modular vs pathway reconstruction of networks. A specific example of network reconstruction, complete with mathematical details would also be nice to see, if such a finished exampled exists.
4. Thoughts
Overall, this is one of the better review articles I’ve read! It really breaks the process into steps, and delves into the issues in each step with just the right level of detail. The figures were wonderful, and really lent themselves to an understanding the broader topic. I also really appreciated the definitions in the side margins on some pages—useful for clearing up misunderstandings that might exist between fields and making sure that the reader was on the “same page”, literally and figuratively.
Colbie,
DeleteI was somewhat confused about these terms at first as well. From my understanding, all three (nodes, modules, and pathways) are useful together, rather than having pros and cons of each. They are distinguished essentially by what the central element is and how that central element is connected to other things. For example, nodular reconstruction looks at a single protein and determines the compounds, reactions, etc. associated with it.
The first paragraph of the section labeled "Scope of reconstructed signalling networks" on page 103 along with Box 2 panel a I think gives the best definitions of the terms. It felt like he focused more on the modules throughout the rest of the paper than the other two after this point.
0 Knew:
ReplyDeleteI was familiar with some of the signaling receptors and pathways discussed but have never studied these in depth. I previously took an immunology course that skimmed the surface of some of the immune pathways the article mentions but did not get deep into details.
Our lab uses some gene expression assays to investigate differentiation of precursor cells on our biomaterials which would be a result of some signaling mechanism; however, we don't generally get into the signaling details. Further, I liked the authors' discussion on "coupling experimental data with .. mathematical analyses." Some of my research uses this approach as well and I think it will provide great insight to the problem and guide future research advances.
1. Learned:
I liked that this article guides the reader through the process of analyzing an interrelated network like the cell signaling network. It seems that the author's approach could be referred to and applied to other types of complex systems.
I also learned from the detail provided on experimental techniques being used for obtaining data needed to model this network. I've used some of these methods previously, but never for these types of experiments.
I had to look up a lot of the receptors and proteins, etc. discussed (although not critical to the main purpose of the article).. so now I am at least familiar with those.
Tables 1 & 2 were also very insightful and I can see myself referencing the articles cited in the tables in the future.
2. Pressing ?:
Modeling a signaling network stoichiometrically seems like the most straight forward approach, but at the same time, is this really feasible? It seems like there are so many feedback loops and causal relationships involved, I can't really understand how this would be developed. The article also mentions the example of making the distinction between NF-kB in the cytoplasm vs in the nucleus.. it seems like including all of these types of factors in a stoichiometric model would be very difficult.
3. Presentation:
I would benefit from a presentation on the difference between using nodes, modules, or pathways for reconstruction of the pathways. I was able to get a pretty good understanding of each reconstruction mechanism, but not necessarily of what each will produce. Could these reconstruction techniques be used together?
4. Thoughts:
Putting numbers to all of these events made modeling these processes seem like a very overwhelming task. While I've been introduced to some of these values in previous bio classes, seeing them all together and their relationships with one another written really drove home the point of why a model is necessary and how useful an integrative model could be to the field.
2) I agree; following any of the 3 main approaches to the stoichiometric level of specificity seems like a very daunting task. I think one way to accomplish it would be to define a very specific network of interest.
DeleteAs the authors indicate, there is a level at which where we define the bounds of the system in a somewhat arbitrary manner. Theoretically, I think that a small, well-defined system could be studied in enough detail to create a stoichiometric model. However, I would like to see an example of such a model in detail.
3)Related to above, I think I would also benefit from detailed examples of the end results of each main pathway.
Kelly McGee
ReplyDeleteClass 02, Assignment 1b
Reconstruction of Cellular Signalling Networks and Analysis of Their Properties
0: What I Knew: I was familiar with the complexity of cell signaling from a physiological standpoint, and was aware of autocrine, paracrine, and endocrine signaling. From other students’ research reports this past summer in the SyBBURE research program, I was aware of studies involving knockdown and fluorescence localization strategies.
1: What I Learned: I gained a more robust appreciation of the complexity of cell signaling and of physiology in general.
I learned about the three main methods (Nodal, Pathway, and Modular methods) for the approach of signal network reconstruction, as well as the levels of specificity of such reconstructions (connective, causal, stoichiometric). I learned that there are a number of various methods being experimented with in attempts to address the biochemical specifics of protein-protein interactions as well as, more broadly, their functional properties.
I learned that the field of cell signaling network reconstruction is at a point where it appears that while we have an appreciation of the incredible scope of data that high throughput genomics and other related technologies have afforded us, both the diverse specificity of protein-level interactions and the incomprehensible magnitude of the calculation power needed to assess all possible interactions simultaneously are daunting challenges.
2: Questions: It is not clear to me what the specific value of the modular method of network reconstruction is. What particular advantage does it afford over the nodal or pathway methods?
P7, Column 1, 2nd Paragraph: “Signaling networks do not function in isolation from metabolic and regulatory processes.” It is abundantly clear that all physiological functions and signals are connected through various pathways. How does one define the limits of a given network of interest?
P8, Column 2, 1st Paragraph: Why are high-throughput protein assays mentioned without any further detail being provided?
3: Presentation Topic: Either a case study use of the modular method of signaling network reconstruction or a thorough review of the processes and experimental limitations of yeast two-hybrid assay studies.
4: Comments: I’m really looking forward to this paper’s discussion.
2) Thought the modular pathway seemed to have some disadvantages to me too, the paper states that the modular pathway helps to place the network spatially and temporally within the whole system. When we further combine the networks together, we need to know if things are occurring in the same space and time to determine their level of interaction. But, I'm still a bit confused as well, so who knows?
Delete0. Knew
ReplyDeleteI knew of the existence of cellular signalling networks as a highly complex description of signalling mechanisms and as a way to characterize mathematical representations of these networks. I am familiar with some of the examples used, for my systems physiology classes have familiarized me with paracrine, autocrine, and endocrine signaling. Other classes have made me aware of the layers of complexity that increase variety of the signalling networks, such as splicing, protein modifications, and ligand-receptor interactions.
1. Learned
It's really important to know the desired complexity of your model, for it is difficult to wholly represent a systemic network with a set of differential equations. Biological context and expected quantitative outcome, as the paper explains, is dictates the type of mathematical formulations of your target network.
In turn, the most labor-intensive step in the construction of a biochemical network is the delineation of known individual biochemical processes that comprise the network. I can imagine just how much work goes into properly fleshing out the network. There are a multitude of data collection methods that I have never heard of like genome-wide location analysis and yeast two-hybrid assays.
2. Pressing
What are the advantages of each type of reconstruction (nodes, modules, pathways)? Do researchers focus on a particular type?
(Page 6, Column 2, Line 10) What is considered an 'entire system'? Does it require every component of the particular signal chain to be considered?
3. Presentation
I would like to talk about the differences between nodes, modules, and pathways, and how they're constructed.
4. Comments
This is a very interesting topic, so I hope we can go into detail about these cellular signalling networks.
2) I think of p53 when I think of the advantages of the nodal approach. It's involved in so many things it would be nice to focus on the p53 node and see where it is connected.
DeleteThough it is not exactly a cell signalling pathway, the first thing that comes to mind when I think of the pathway approach is the coagulation cascade, a sequential cascade which doesn't really have much of a practical effect in the middle, but only comes to fruition at the end. Signalling pathways of the same type would benefit, I think, from pathway modeling.
I'm also not sure of the benefits of the modular pathway. They go into one surface level example in the text but I wasn't satisfied with that.
This comment has been removed by the author.
ReplyDeleteNatalie Hawken
ReplyDeleteReading Assn 02: Reconstruction of Cellular Signaling Networks and Analysis of Their Properties
0. KNEW
Through my coursework in biology and biochemistry, I have learned about signaling pathways mostly as discrete units, not as complex interconnected networks. I was familiar with autocrine, paracrine, and endocrine signaling (especially in regard to the immune response), and I somewhat understood the timescale associated with these processes. I had heard of orders of magnitude in many different science courses, but I had only applied it to signal transduction pathways within cells. I did know a good amount about the types of modifications (alternative splice variants, post-translational modifications, dimerization, ligand states) that lead to the huge number of components, but I had no idea about the extent of variations made possible by the combination of all of these factors. I had heard about high-throughput techniques, but only in regards to chemical compound testing.
1. LEARNED
This paper was insanely informative to me in terms of exposing me to how all of the variations in signaling molecules and pathways lead to what can seem to be infinite responses and connections to the overall signaling network. I began to understand the magnitude of the complexity when the various combinations in proteins accounted for "nearly a 20-fold increase in the number of protein states over the respective number of genes" (pg 101). Also, since I had never really thought about the connections between membrane receptor proteins, I didn't realize how the ligand-bound vs. ligand-unbound states of each type of receptor could create so many unique combinations of signaling into the cell. I found the figure detailing how they estimated the number of unique receptors on a cell particularly informative. Because they compiled info about cell size, membrane composition, receptor protein area, etc, they were able to use very simple calculations to find their estimate.
I found the processes to reconstruct the networks very interesting. Since I'm used to studying a very small part of a very specific pathway, I was unaware of how scientists compiled data from so many studies on different topics in order to make their findings. Though I had experience with most of the bottom level points in Fig. 1, this article was my first exposure to the process of interweaving these studies into a whole new conclusion.
2. MOST PRESSING QUESTIONS
This paper gives a lot of estimated numbers of variations and combinations, derived by assuming in vivo numbers fit into mathematical equations (see pg. 102 for ligand equation, pg. 101 for post-translational modification assumptions). Since this paper is over 10 years old, hopefully more studies have be able to show specifically how many of these combinations exist. So, how closely do Papin's calculations match actual counts? If there are discrepancies between the two, what did Papin overlook/model incorrectly in order to get his result?
When discussing limitations, Papin states "modular descriptions of signaling networks require intuitive, subjective, human-based decisions" (pg. 104). How to do the benefits of the modular style outweigh the negatives that prevent the automation of the process. To me, it seems like a process that requires little to no human decision-making would be overall the best way to make the most integrated signaling network in the fastest time.
Continued...
...Continued
ReplyDelete3. PRESENTATION TOPIC
I was most confused by the section "Data Collection for the network reconstruction process" (start pg 105). I think a presentation on how each of the techniques (chip assays, yeast two-hybrid assays, ICATs, RNAi, etc.) are used, what materials they require, what information they produce. This would help to understand how they get all of the necessary information to create the network. Also, I would be interested on a presentation about the mathematical analysis (what programs, how long does it take, what information do they find out) used on the networks.
4. THOUGHTS
I was very concerned at first that this paper was going to be way too dense for me to understand, but I was happy that I was actually able to learn a lot about the process to create a quantitative model. Though it would be interesting to look into the specific of how the models are made and tested, I enjoyed getting a general overview. As someone who researches just a single type of glutamate receptor, I sometimes forget how massive the field of signaling becomes when all of the components are viewed and studied together.
Part 1/2
ReplyDeleteSylvia Morrow
Asgn1B_Reconstruction of Cellular Signalling Networks and Analysis of Their Properties
0. KNEW: Again, very little about the specific biological processes; however, the general concepts of networks are somewhat familiar to me in a more physics/math context.
1. LEARNED:
Some more general biology concepts e.g. transcription factor, phenotype, ligands.
Experimental biology has used a vast array of techniques to approach problems from many different angles depending on what information is most relevant to a specific study. Network reconstruction faces the challenge of bring this data together to extract the big picture. My understanding was that because different techniques were used, each different set of data must be very carefully entered into the network in a way that very specifically states what limitations there are on that experimental technique/what really does the data say as opposed to just a set of possible inferences.
A few concepts I found particularly interesting:
--"cells can express a few receptor types (~10–40) in high numbers (~105 per cell), or many receptor types (~ 2,000–30,000) in small numbers (~102 per cell) on their surface. The former might be more common in highly differentiated and specialized cells, whereas the latter might be more typical of stem cells or undifferentiated cells." (103) I'm not 100% sure that I'm interpreting this correctly, but this seems to say that given a constant surface area availability the degree of specialization of a cell is essentially a measure of if it chooses to detect a few types of signals with high precision or lots of signals with low precision.
--the enormous complexity of a biological system: A huge amount of physics knowledge begins with a "simple model" that makes a bunch of assumptions and approximations but still yields a reasonable answer in certain scopes. Based on this article it seems like it is near impossible to start with some relatively simple, isolated module and still get biologically-relavent results. It seems like the best we have to hope for at this point is to at least compress some complex things into simpler approximate inputs.
-- Drosophila melanogaster is the scientific name of the common fruit fly.
2. PRESSING ?:
--How common is it for a biological system to be in an unstable equilibrium (so that a small perturbation in input leads to a significant change in output)?
--pg 102 talks about combinatorics which was familiar to me from statistical mechanics, but in thermodynamics you typically end up with combinations that while technically unique are not unique in a physical sense (an corrections are made for this). this doesn't seem to be the case in this discussion, so I would like to have a better understanding of why this is the case.
--There seems to be a subtle difference between "assay" and something like experiment/analysis but I can't quite figure out what it is.
part 2/2
DeleteSylvia Morrow
Asgn1B_Reconstruction of Cellular Signalling Networks and Analysis of Their Properties
3. PRESENTATION:
--(as mentioned by a few people already) if someone could present a case study that would probably be really informative.
--also as mention already, more detailed overview of the data collection process
4. THOUGHTS: I got increasingly excited about the article as it went on. In particular, discussion about dynamic analyses and perturbative approaches were nice because they were very familiar and easy to understand but also because it's pretty clear that this field is still in the stages of fairly naive modeling, and I think it's really interesting to think about the levels of complexities that will be added once there is a better general foundation. I would think that a long-term goal of this type of work would be to not just do better reconstructions of these modular networks but to start piecing together networks to achieve a comprehensive (albeit enormous, though it would probably be easy to have off switches for irrelevant modes) model of cellular signaling networks.
Also, son of sevenless was such an interesting name that I looked up the origin and was delighted to find that there is also a bride of sevenless aka BOSS.
An assay is a type of experiment! Not all experiments are assays, but all assays are experiments. Generally assays are looking at the amount of presence of certain activity (for example, in previous research I did, that activity was contraction of actomyosin networks).
DeleteDanny McClanahan
ReplyDeleteClass 2 Article 2 (Reconstruction of cellular signalling networks and analysis of their properties)
0. Knew:
I'm currently working on a project in the iGEM lab modelling dynamic behavior of signalling networks, so I was somewhat familiar with the mathematical representation of signalling. The final small section on dynamic behavior of networks with Monte Carlo sampling describes part of our approach to the project, for example.
1. Learned:
I was far less familiar with what types of chemical/biological factors end up producing this sort of behavior, and how that integrates into the rest of the research process. Our project is purely abstract and mathematical right now, so the bridge to concrete reactions which cause these behaviors, as well as learning behavior from causal reconstruction, was all new to me.
2. Questions:
What methods are used to aggregate data of different types, for example connectivity vs causal reconstruction? What about data at different scales? How is uncertainty in causal reconstruction represented in a model? How can interactions be tested both in isolation and in vivo (by this I mean how can "true" interactions be teased out from the maelstrom of reactions in a true cell, and how can we be sure measured interactions are actually possible to create in a real cell at the same time)?
3. Presentation Topic:
The discussion of the scale of a model, both in time and space and number of individual interactions (e.g. in a signalling network) is important and also potentially confusing. The utility of large vs small-scale models and experiments differ. In addition, the iterative approach to hypothesis generation and then experimentation and back makes use of data at all scales. Discussion of what these different scales mean and how they can be used together is complex and important.
4. Thoughts:
The article seemed far more approachable than the previous one, often explaining any new or complex ideas with images, even, or at least a glossary in the margin. This is likely made possible by the high level of abstraction the article remained at -- few descriptions of actual data, experiments, or mathematical models were actually given in the article, relying instead upon citations to give more information (which is fine). The article mostly focused on general guidelines for how to perform effective modelling of signalling networks for the goal of hypothesis generation as opposed to specific methods of doing so.
0.Know: As stated in the text, the article contained information that focused mainly on cellular signaling networks that affect transcriptional regulation. I know the basics of transcription and its role in ultimately generating proteins from genes, so I was familiar with some of the concepts mentioned. I know that that exons are the DNA regions that get coded to RNA (while introns, not mentioned, are the regions of the genes that do not get coded). I am generally familiar with the process of transcription and with the ensuing translation process. Additionally, I know that the extracellular matrix is essential regarding cell recognition of other cells and of its environment.
ReplyDelete1.Learned: From what I knew about cellular signaling networks, I thought of it regarding the maintenance of an organism; however, I did not stop to think that it is also very much applicable to the formation of an organism and the cell differentiation that follows. One of the things involved in the reconstruction of a cellular network is the estimation of the number of components in the network; signaling networks are so complex because the number and type of components (and the interactions between components) is so vast. For example, mentioned in the article was that there are several different types of receptors and that they may be expressed in either large or small numbers depending upon the cell. I learned about considerations that define the orders of magnitude, and I was exposed for the first time to the different ways (nodal, modular and with pathways) one can approach the reconstruction of a signaling network. I also found it interesting that the interactions between networks can be discussed in the perspective of a single cell that transmits and receives a signal once the other considerations for reconstructing a cellular network are taken into account.
2.Pressing Questions: The article mentions near the beginning that many other processes are also tightly coupled with cellular signaling networks. How would network reconstruction for cytoskeletal organization, metabolism, and the other processes mentioned differ significantly from the cellular signaling networks that affect transcriptional regulation? Also, I read that there are proteins that arise from differential splicing that have different “proteolytic degradation properties.” What may these properties be? Furthermore, I think the three approaches to signaling network reconstruction (with nodes, with modules, and with pathways) could have been elucidated more.
3.Presentation: I think it would be helpful to see an example of a mathematical analysis derived from the different considerations of a signaling network. I know that a mathematical analysis in conjunction with the qualitative description of a signaling network is useful, but since I have not seen how these work together or how one helps provide information about the other, it seems like an ominous, vague truth. Brief examples of different networks could also be helpful.
4.Thoughts: I did like that the article provided numbers and scales for reference, but it did become overwhelming at times (not that I do not wish to further understand how huge the genome and proteome are). I like how the article was organized with different topics; while the content was not always easy to follow, the structure was.
I will work on responding to these comments later today - I'm at a 3D tissues conference with lots of Pharma representation, and there are good discussions of just these topics. See you all Thursday.
ReplyDelete0. KNEW:
ReplyDeleteI knew about basic cell signalling including autocrine, paracrine, hormonal, and neurotransmitter responses, the timing of these signalling methods, and mechanisms of action. I knew about post-transcriptional modifications and what exons are. I knew nothing about network modules. I also knew nothing about the experimental methods used for collecting network data.
1. LEARNED:
I learned the importance of scale in cell signalling, something that I had been aware of but not ever thought about. I was not aware of how many different receptors and kinases humans possess. I actually would have thought we have more kinases. I wasn’t aware that genes are spliced into multiple different functional genes and how this was important to signalling. I learned the importance of considering the combinations of different entities in a network as this affects what the network is capable of a whole. I learned that adding more states to a network can greatly increase the number of pathways within the network. I didn’t realize just how many receptors cells could possess. I learned what exactly a “yeast two-hybrid experiment” is.
2. PRESSING ?:
<< BE SURE TO LIST PAGE/LINE NUMBERS>>
Heading: Signal transmission -> “An appreciation of the orders of magnitude in signalling networks is the first step in analysing their properties because it allows for reasonable assumptions regarding limitations on possible network function”, What limits on the network would this be? Is this simply the reach and speed of connections?
What visual networks were constructed before this paper was released? What technology existed to aid these processes and how does that differ from what is available today?
3. PRESENTATION:
The paper mentions “high-throughput techniques [that] elucidate the mechanisms that connect an extracellular signalling stimulus to the control of transcription factors” and how this technology is new… in 2005. An interesting topic would be what this technology looks like today and how it conforms to the challenges the paper presents.
4. THOUGHTS:
I think this paper, due to its graphical nature, will make an excellent paper for presenting.
0. Knew
ReplyDeleteI knew of the existence of cellular signalling networks as a highly complex description of signalling mechanisms and as a way to characterize mathematical representations of these networks. I am familiar with some of the examples used, for my systems physiology classes have familiarized me with paracrine, autocrine, and endocrine signaling. Other classes have made me aware of the layers of complexity that increase variety of the signalling networks, such as splicing, protein modifications, and ligand-receptor interactions.
1. Learned
It's really important to know the desired complexity of your model, for it is difficult to wholly represent a systemic network with a set of differential equations. Biological context and expected quantitative outcome, as the paper explains, is dictates the type of mathematical formulations of your target network.
In turn, the most labor-intensive step in the construction of a biochemical network is the delineation of known individual biochemical processes that comprise the network. I can imagine just how much work goes into properly fleshing out the network. There are a multitude of data collection methods that I have never heard of like genome-wide location analysis and yeast two-hybrid assays.
2. Pressing
What are the advantages of each type of reconstruction (nodes, modules, pathways)? Do researchers focus on a particular type?
(Page 6, Column 2, Line 10) What is considered an 'entire system'? Does it require every component of the particular signal chain to be considered?
3. Presentation
I would like to talk about the differences between nodes, modules, and pathways, and how they're constructed.
4. Comments
This is a very interesting topic, so I hope we can go into detail about these cellular signalling networks.
Benjamin Reisman
ReplyDeleteClass 2 Article 2 (Reconstruction of cellular signalling networks and analysis of their properties)
0 Knew
I was somewhat familiar with the content of this paper, though less familiar than I was with the cisplatin paper. My undergrad lab studied was a networks lab and there was a subgroup working on biological networks. My project was more focused on bibliometrics (measuring collaboration and impact in science using citation data), but I was exposed to network construction and structural network analysis.
At Vanderbilt I rotated with Gregor Neuert briefly working on MAPK signaling in yeast, and it’s exciting to know that faculty at Vanderbilt are actively leading this field.
1 Learned
Before reading this paper, I hadn’t considered how the various scales involved in biological systems would present a problem for network approaches to cell signaling. It was also interesting to note how different groups have attempted to solve the scale problem, in particular those that integrate different types of omics data together to create unified models of signaling, metabolism, and gene transcription.
2 Pressing Questions
Can we model biology in such a way that a model constructed at one scale can provide insights at the next level up? Such a model would display emergent properties. The last paragraph on page 104/first paragraph on page 105 provides one such example, in the case of β adrenergic signaling in rat cardiomyocytes. In chemistry/physics, statistical mechanics provides the link between quantum mechanics (small scale) and thermodynamics (large scale). Is such a unifying model of biology possible through systems biology?
On a more concrete note, I’m curious about the implications of the structural network analysis. From the few papers we’ve read so far, it seems like the most important step in any project is to choose the right “sized” question, in this case with regards to scale. If trying to construct a causal model of signaling form omics data is a physicists attempt at reductionist biology, how should we consider structural network analysis where the object of study is analyzed as a whole?
3 Presentation Topic
I’d like to review the different classes of networks and (ideally) how they apply to biology. What is a scale free network and why should we care? What is a small world network? Etc…
4 Thoughts on systems biology
If causal models from systems approaches need to be confirmed using traditional approaches to molecular biology, do network models require the same confirmation or can they stand on their own?
Very neat paper, looking forward to the discussion.
0 Knew
ReplyDeleteThe topic of this paper is incredibly broad. Necessarily, the author reviews major characteristics that any successful model of signaling networks must have and discusses reviews both experimental obstacles and possibly unexpected conceptual obstructions to this characterizing process. I began reading this article with the awareness that sub-systems with different intrinsic spatiotemporal scales must in any good systems model. Then, the backdrop of vandy’s work in RTA (= Mass spec+statistics) provided me with a working experimental model with which to compare other efforts.
1 Learned
If we are able to set up detailed mass spec analyses of systems that can measure the concentrations of essentially every biologically significant compound within a system, then, (I thought) what else could you do to give yourself more mineable data besides improving the spectroscopy? This is my RTA hubris speaking. After reading this article, I was finally able to sweep that naïve view away and to return to what may seem obvious- different experiments allow you to collect data that may be more interesting that concentrations! So, with RTA as working model, I say that we simply couple it to more tests. Then, the problem of reading the implications of the intertwined data sets becomes interesting.
2 Pressing Questions
how can we couple various experiemtnal techniques simultaneously? Some are light based and don’t really intereact with cells, some depend on the destruction of cells. We may be able to determine the sets of tests comprising of groups whose members are compatibile with any other member of the group.
3 Presentation Topic
Compatible experimental techniques
4 Thoughts on systems biology
My human insight into problems is uncharacteristically unhelpful in the domain of elucidating signaling pathways.
Class 02, Assignment 1B
ReplyDeleteReconstruction of cellular signalling networks and analysis of their properties
0. KNEW:
I was familiar with autocrine, paracrine and endocrine signalling pathways and the diverse spatiotemporal scales cellular signalling, as well as some of the specific signalling pathways and proteins (MAPK, NF-kB, RAS, GPCRs) used to illustrate various concepts.
1. LEARNED:
Scale of specificity (20-fold increase) that is achieved via transcript PTMs: 30,864, 10,360 and 3,000 distinct receptor, kinase and phosphatase states. I also learned about ICAT and SILAC for the first time.
I learned of the various approaches that are used to reconstruct signalling networks; i. nodal identification, ii. pathway identification, and iii. signalling ‘module’ identification. I also learned how each approach is used to gain better contextual specificity, because while none are complete all are necessary to integrate disparate data sources upon which this is based.
I learned that the limiting factor in signalling network analysis now appears to be shifting from biological experimentation/data collection to mathematical modeling/computational capabilities, or maybe now the latter is emerging because the former has advanced to the point where it has reached a precipitating event.
2. QUESTIONS:
(Page 109, Column 1, Line 6) “Recent efforts to automate network reconstruction processes and extract molecular-interaction data from the literature are important computational developments that will allow these more extensive network reconstructions.” How has the field developed since the publication of this article? Have stoichiometric signalling network reconstructions proven to have system-level predictive capability? How would you verify causality?
3. PRESENTATION TOPIC
Approaches to hierarchical organization in systems biology; While the author specified nodal, pathway and signalling ‘module’ identification techniques it would be nice to to go into a little more detail on each.
4. THOUGHTS
The article provided a huge breadth of information but the article structure, figures and footnotes made it much more accessible. I’m interested in learning more about the mathematical modeling the author describes at the end of the article.
0. KNEW:
ReplyDeleteIn previous biology classes I learned about different types of signalling networks and the various timeframes in which they operate. I understand the basics of how these signals effect changes within the body. I was also familiar with the concept of alternative splicing and the variations that arise due to this phenomenon.
1. LEARNED:
I learned the complexity that goes into these signaling networks is unlike anything I would have previously imagined. The orders of magnitude that are involved in these networks is remarkable, but is easier to grasp when you understand the scale of the human body (~25,0000 genes, 10^14 cells, 200 cell types, etc.). The fact that stood out the most was that the number of receptor proteins for a somatic cell is estimated in the millions. After grasping the large numbers that involved I learned why computer and mathematical models would be used to analyze these systems. I also learned how much human decision making goes into setting the specific parameters of these models.
2. PRESSING QUESTIONS:
How can these models be improved to require less human input? Are there any restrictions on these models because of the technology itself? Would advances in computers allow for a more rapid analysis since the number of iterations that are required are so large.
3. PRESENTATION TOPIC:
Specific examples of what has been (and can be) accomplished by these models would give me a better understanding of their potentials and limitations.
4. THOUGHTS:
I'm excited to see how the presentation captures the scale of what is required for these signaling networks. I hope someone with some experience in modeling can provide insight as to what they do.
Stephen Lee
ReplyDeleteClass 03, Assignment 1b
Reconstruction of Cellular Signalling Networks and Analysis of Their Properties
(0) Knew:
I knew about the diversity of autocrine, paracrine, and endocrine signaling events prior to reading this paper. I’m also familiar with a handful of cell signaling mechanisms, like GPCRs and tyrosine kinases. I am, however, unfamiliar with the reconstruction of cell signaling networks and whole-network analyses as described.
(1) Learned:
I learned that cell signaling networks operate over many orders of magnitude even at the cell surface level; signals may be mediated by thousands or tens of thousands of surface structures of diverse conformations. I also learned that identical signals are transmitted through a multitude of pathways from initial input to final output to mediate amplification and feedback, as opposed to the more simplistic notion that signaling proceeds uniquely as a linear cascade. I also became familiar with the iterative process of network reconstruction.
(2) Pressing Questions:
(Page 2, Column 1, Line 1) What challenges are posed by processes like cytoskeletal organization, organelle assembly, vesicular trafficking, and metabolism (as mentioned in the paper) that impede their mathematical formulation? How may they be resolved?
(Page 10, Column 1, Line 3) How is spectral analysis used to define protein-protein interactions in the yeast proteome? What makes this (these) method(s) more reliable than a two-hybrid assay, for example?
(3) Presentation Topic:
It may be useful to explore in greater detail the mapping of cellular signaling network reconstruction, such as that depicted in Fig. 1. This figure seems to summarize the gist of the paper well, but could be accompanied by a more thorough description of the “existing biological knowledge” it references (processes like ChIP, perturbation analysis, and cDNA assays).
(4) Thoughts:
This paper did well to explain a complex and specialized effort in a way that a non-specialist could understand. Curious as to how the conversation will develop in class, and how others benefited from the reading.