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.
Saturday, November 12, 2016
SysBio16 Asgn_24A_Class_24_Article_19_2016_11_17
CN-Bio Networks: Read-Scan Article 19 LG
Alexopoulos, et al., Networks inferred from biochemical data reveal
profound differences in Toll-like receptor and inflammatory signaling
between normal and transformed hepatocytes. Mol.Cell.Proteomics 9.9
(2010). Reading team - read carefully; Class - read abstract, captions,
conclusions so you know what's going on and can ask questions of our
guests. Post a PCRC on the blog.
Subscribe to:
Post Comments (Atom)
Colbie Chinowsky
ReplyDeleteAssignment 24A, article 19—Networks inferred from…
1. Knew
In other one of my classes we recently had a lecture on HTS, and it’s advantages, so I was somewhat familiar with the techniques used in this paper.
2. Learned
This paper focuses on the interference of function relationships via high throughput screening data. Macromolecules are represented by nodes, and the interactions as edges. Hepatocytes are used as they were expected to be physiologically informative, as the liver has many different functions that are carried about by the hepatocytes. Once HTS was completed, comparative regression analysis was used to reveal the differences between transformed and wild type hepatocytes. xMAP assays explore the levels of cytokines. Cells were exposed to seven different extracellular ligand “cues”, including all mediators, inflammatory factors, mitogenic factors, a ErbB1/2 ligand and the TLR4 agonist LPS. Additional data was collected by treating cells with seven different small molecule kinase inhibitors, which produce ~90% target inhibition in HepG2 cells. The primary data was analyzed to find differences between primary hepatocytes and HepG2 cells using MLR. These differences were evident in the differences between regression weights for induced secretion of cytokines, and factors for mitogenesis and cell survival. Additionally, they found cell-type specific protein signaling networks, and worked to recreate these networks.
3. Pressing
Has any work been done integrating the ECM into system biology and signaling pathways?
4. Presentation
My memories of classes are starting to blend together, so I can’t remember if it’s this class that’s already done a short presentation on HTS in general. If not, we should.
5. Thoughts
This paper covers a lot. It covers all of it pretty well, considering the amount of text given to each topic, and enhances the idea that HTS and MLR can produce a vast amount of data. It did make the paper harder to get through.
0. Knew:
ReplyDeleteI was slightly familiar with different aspects/properties of the liver and its functions from discussions in class. I also had a familiarity with network construction from correlation (regression) weights.
1. Learned:
This paper looks at combining the large data driven analysis of signaling networks with the functional analysis of these networks through perturbations to compare the signaling networks of human hepatocyte cells and liver cancer cells. Specifically, this shed light on the difference between these two cell types in terms of how they respond inflammation.
The main thing I learned from this paper is that it is possible to acquire a lot of data without actually acquiring a lot of data. This is seen in the discussion section of the paper when the authors state that the success of this type of data analysis to produce useful network inference was most likely due to the fact that there were multiple perturbations performed on the cell lines. Therefore, even though there were not a lot of data points for each cell state, the fact that there were a lot of cell states from the multiple perturbations still gave enough information for the network inference.
2. Pressing Questions:
Could using some other type of network inference on the same data produce any new information? Is there any way to know if the type of inference method chose gives the “best” or “most” information?
3. Presentation Topic:
The end of the paper gives somewhat of a comparison of the methods of this paper to other network inference methods. It would be interesting to see which types of network inference one would decide to use based on various conditions such as amount of data one is able to collect, signal-to-noise ratios, and number of genes, proteins, etc. in the network.
4. Thoughts:
While some parts of the paper were hard to get through, it was nice seeing an entire process of going from data to networks and seeing what steps/assumptions are made to produce these networks.
0. KNEW:
ReplyDeleteI knew about how biological networks are typically represented and what information they fail to give. I knew about multilinear regression. I knew how the immune response works. I was aware of how cell signalling works. I knew about flow cytometry.
1. LEARNED:
I learned what xMAP is. I learned about how to control for variability and purity of cell lines. I learned how exactly this group went about perturbing the cell lines with various signaling molecules and kinase inhibitors. I learned about the concept of overdetermination. I learned an example of a signature response set of genes. I learned the acronym PIN.
2. PRESSING ?:
How would this have worked out with Spearman coefficients?
Is RTA using any of the methods for generating data such as RNAi screening?
3. PRESENTATION:
I would lay the background of this presentation with the differences of this and what we are trying to accomplish (i.e., RTA). These differences are the fact that this group set out with known pathways and cell perturbation processes. They also measured far fewer intracellular signals I believe. I would emphasize the linear approach used in analysis. This group used weighted edges for their graphs which by looking at figure 5, highlighted downstream genes. This makes sense and can be explained using the figure. Again, the highlighted genes are known.
The node-edge graphs of figure 4 could be compared to RTA figures and emphasized that theses are simply correlation. Directions were defined through the known literature, not through statistical determination.
I think the differences in intermediate cell signalling is noteworthy and could be used as a segway to bring in the power of these networks, but this is seemingly heavily dependent on the cell lines chosen.
Would also mention the use of only biochemical data as emphasized in the discussion and signal-to-noise ratio of their approach vs genome-scaling approaches (RTA); they got by with using known interactions and relations.
4. THOUGHTS:
This paper was straightforward and emphasized the relevant points by maintaining minimal discussion of lab procedures. Very grounded paper and should make for a great discussion.
Sylvia Morrow
ReplyDeleteAsgn24A_LG Alexopoulos, et al., Networks inferred from biochemical data reveal profound differences in Toll-like receptor and inflammatory signaling between normal and transformed hepatocytes.
0. KNEW: Basics of networks as discussed in class, and a little about high throughput screening which has come up in class discussions but which we haven't gone into detail on.
1. LEARNED: It has bothered me the whole time that we've been talking about networks that the networks often taken data from many different cell types and put them together in one figure with no obvious justification for doing so. This paper emphasized the problems with that approach and chose to create "smaller networks that are cell-specific and focused on the receptor-mediated signaling processes known to play an important role in oncogenic transformation" (1858). The motivation for this paper was that "how diverse signaling proteins are coordinated in normal and diseased liver cells remains poorly understood" (1850) which directly impacts our ability to treat liver diseases and cancers. The study used comparative regression analysis of high throughput data. A few sections of the experimental procedure were technical enough that I had a hard time extracting useful information ("Cell Culture and Signaling Experiments," "xMAP Assays," "Reagents"). This paper addressed sources of uncertainty through additional assays, namely to evaluate the effects of "experiment-experiment and donor-donor variability and the purity of the primary cell preparations" (1851). The experiment design was influenced by information about networks in the liver found in the literature. This study found that primary cells respond more to inflammatory cytokines whereas transformed cells respond more to growth factors, and even when there is overlap in response type it occurs at different levels or on different time scales.
Also in general I didn't know much about the liver so it was interesting to get an overview of its function.
2. PRESSING ?:
--Many of the figures have an enormous amount of information on them. I think it will be helpful to talk through these in class.
3. PRESENTATION:
--May be useful to do a specific presentation on high throughput screening to get a better understanding of the pros/cons.
4. THOUGHTS: A dense but informative paper.
This comment has been removed by the author.
ReplyDelete0. KNEW
ReplyDeleteI was somewhat familiar with the function of the liver and basics of the ontology process in hepatocytes. Also, from previous papers in class and my presentation on hierarchical clustering in weighted gene network correlation analysis, I had a base understanding of network construction without a priori defined modules/networks.
1. LEARNED
The main point of this paper was to define gene network differences in normal hepatocyte cells vs cancerous liver HG2 cells. This was done through extensive perturbations to the existing gene network in the cells. I also learned about xMAP techniques and more about practical biological cell cultures, as opposed to just the data portion of the paper. I feel another major point of the paper provided a discrepancy between primary cells and cancer cells in terms of perturbation of certain classes of molecules. In other words, the hepatocyte cells, according to the data, were more greatly affected by inflammatory cytokines (greater differences in downstream pathways), while HG2 cells were more greatly affected by certain growth factors. In addition, this paper makes apparent the integration of network inference "from scratch" with existing canonical data in order to limit the proteins and genes of interest so as to efficiently cut down on relevant output data.
2. PRESSING ?
Have there been practical applications of the finding of this paper after the fact?
3. PRESENTATION
More on xMAP and cell cultures (a bit of breakout from data analysis)
4. THOUGHTS
dense, but able to be followed closely through all the detail. The discussion section was very helpful in reiterating major conclusions that were not readily apparent in the experimental and results sections.