BIOL398-03/S13:Class Journal Week 13

(Difference between revisions)
Jump to: navigation, search
 Revision as of 01:33, 19 April 2013 (view source) (Answered #3)← Previous diff Current revision (02:03, 19 April 2013) (view source) (→Helena M. Olivieri|Helena M. Olivieri Week 13 Journal) (2 intermediate revisions not shown.) Line 17: Line 17: #*We can use a line of best fit in the ter Schure et all data to use the least squares technique. #*We can use a line of best fit in the ter Schure et all data to use the least squares technique. #Read over the Vu and Vhoradsky paper in this week's reading list. Is the method of the paper different from that in class? If so, in what ways? #Read over the Vu and Vhoradsky paper in this week's reading list. Is the method of the paper different from that in class? If so, in what ways? - #*Although the Vu and Vhoradsky paper used a nonlinear differential equation model of gene expression and our class used a similar sigmoidal model, we did not perform a polynomial fit. Vu and Vhoradsky used a polynomial fit as an approximation of real expression profiles because the true profiles are obscured by experimental errors. + #*Although the Vu and Vhoradsky paper used a nonlinear differential equation model of gene expression and our class used a similar sigmoidal model, we did not perform a polynomial fit. Vu and Vhoradsky used a polynomial fit as an approximation of real expression profiles because the true profiles are obscured by experimental errors. Moreover, they used more genes. + [[User:Laura Terada|Laura Terada]] 02:00, 19 April 2013 (EDT) ==Week 13 Ashley Rhoades== ==Week 13 Ashley Rhoades== Line 111: Line 112: #*The V&V paper used a similar approach, including solving their differential equation using ode45 in Matlab.  There were a few differences.  They had many more data (time) points to fit, and used a 6-degree polynomial  to approximate the expression of the genes.  (6! ack!  That is an extremely complex polynomial on such a relatively small number of data points.) #*The V&V paper used a similar approach, including solving their differential equation using ode45 in Matlab.  There were a few differences.  They had many more data (time) points to fit, and used a 6-degree polynomial  to approximate the expression of the genes.  (6! ack!  That is an extremely complex polynomial on such a relatively small number of data points.) [[User:Elizabeth Polidan|Elizabeth Polidan]] 01:33, 19 April 2013 (EDT) [[User:Elizabeth Polidan|Elizabeth Polidan]] 01:33, 19 April 2013 (EDT) + + ==Helena M. Olivieri| Helena M. Olivieri Week 13 Journal== + [[User:Helena M. Olivieri|  Helena M. Olivieri]] + #How might you change the network to add more genes/transcription factors? + #*Because last week we assessed which unique genes we would like to add to YEASTRACT model, it will soon be necessary to add actually add our genes to our Excel matrix. It will be necessary to, thus, continue to use YEASTRACT to generate a matrix that will make connections between the new unique genes and the previously studied genes. Results from already produced matrices may be inserted into our current Excel file. + #How might you run a computer experiment to examine the deletion of a gene? + #*Removing the gene completely or manipulating the amount of gene regulation in the excel file could emulate a gene deletion. + #In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments? + #*Least squares is a standard approach of approximating a solution. Data fitting can be identified as the most important application in terms of the least squares technique. Thus, because the ter Schure experiments tend to present linear data it would be appropriate to determine the parameters by using a linear best fit trend line. + #Read over the Vu and Vhoradsky paper in this week's reading list. Is the method of the paper different from that in class? If so, in what ways? + #*There were really only two differences I was able to note between the method of the paper and that of our class. The differences I saw were the use of a nonlinear model and a greater quantity of studied genes.

Reflection

• By adding more rows and columns to the excel sheet with transcription factors and genes to compare to the other ones being tested.
• If the gene has no affect on any transcription factors, it is as if it were deleted, so one could put zeros everywhere it is affecting expression (has a one).
• The least squares technique minimizes the sum of the squares of the residuals or error produced by the model in relation to the data. Using the best fit line from the ter Shure data, we could change our parameters in the model we ran until our best fit lines start looking like those in the ter Shure paper.
• I noticed they were looking at change in gene expression during the cell cycle and other developmental processes, not response to stimuli or cold shock as we are looking at. They also did not group genes based on similar expression patterns, but looked for transcription regulatory networks by looking at which transcription factors bound to which promoters. They fit each gene profile with a polynomial degree n

Kevin Matthew McKay 01:44, 17 April 2013 (EDT)

Laura Terada

1. Look over the excel workbook input file. How might you change the network to add more genes/transcription factors?
• In the Excel spreadsheet, you can add more columns/rows to add more genes/transcription factors to run.
2. How might you run a computer experiment to examine the deletion of a gene?
• Instead of 1, you could input a value of 0 to show that there is a deletion. This change will show how the deletion affects other genes/transcription factors.
3. Return to the chemostat. In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments?
• We can use a line of best fit in the ter Schure et all data to use the least squares technique.
4. Read over the Vu and Vhoradsky paper in this week's reading list. Is the method of the paper different from that in class? If so, in what ways?
• Although the Vu and Vhoradsky paper used a nonlinear differential equation model of gene expression and our class used a similar sigmoidal model, we did not perform a polynomial fit. Vu and Vhoradsky used a polynomial fit as an approximation of real expression profiles because the true profiles are obscured by experimental errors. Moreover, they used more genes.

Laura Terada 02:00, 19 April 2013 (EDT)

Week 13 Ashley Rhoades

• How might you change the network to add more genes/transcription factors?
• Add more transcription factors
• How might you run a computer experiment to examine the deletion of a gene?
• Delete a gene or alter its transcription to zero to look at the effect on the other genes.
• Return to the chemostat. In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments?
• You could look a best fit line of the ter Schure data to estimate the parameters
• Read over the Vu and Vhoradsky paper in this week's reading list.Is the method of the paper different from that in class? If so, in what ways?
• Vu and Vhoradsky set limit of parameters to 500.They looked at cyclin networks.

Salman Ahmad

1. Look over the excel workbook input file. How might you change the network to add more genes/transcription factors?
• More genes and transcription factors can be added to the excel file very easily. All you would have to do is add more columns and rows to the table and add the new genes and transcription factors.
2. How might you run a computer experiment to examine the deletion of a gene?
• If a gene is deleted it will have no ability to up or down regulate any other gene. The easiest way to examine the deletion of a gene is to delete it from the model.
3. Return to the chemostat. In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments?
• You would have to use the least square technique to fit a line through their data. If you are able to get the lines to be similar, then you would have an estimate of what the parameters were for their experiment.
4. Read over the Vu and Vhoradsky paper in this week's reading list. Is the method of the paper different from that in class? If so, in what ways?
• In the Vu and Vhoradsky paper they look at 40 cell cycle regulated genes. The method for calculating is not that different from the one we went over in class. In this paper the methods are being used on a greater scale. "The procedure was applied to 40 yeast cell cycle regulated target genes and 184 potential regulators" In the paper there were also comparisons between linear and non-linear modeling.

Salman Ahmad 18:13, 18 April 2013 (EDT)

James P. McDonald

1. How might you change the network to add more genes/transcription factors?
• You can add more genes/transcription factors in the network in the excel file by adding additional rows and columns with new genes or transcription factors.
2. How might you run a computer experiment to examine the deletion of a gene?
• If you add a 0 in place of a 1 it represents the deletion of that particular gene. You could also delete the gene from the excel file. You can then run the experiment with this deletion to see its effect.
3. Return to the chemostat. In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments?
• The least squares technique relies on an approximation and data fitting, therefore placing a line of best fit into the ter Schure data can allow you to approximate the parameters of the experiment.
4. Read over the Vu and Vhoradsky paper in this week's reading list. Is the method of the paper different from that in class? If so, in what ways?
• They used a similar method and analysis as we did, but they they used a polynomial fit, which differed from our model. This was done as an approximation. They also had a larger number of genes that they analyzed.

Matthew E. Jurek Week 13

1. How might you change the network to add more genes/ transcription factors?
• Adding more transcription factors relates to last weeks assignment, when additional transcription factors were added to the network before generating a network map. Likewise, YEASTRACT could be used to explore additional transcription factors. These factors could be added to the matrix, and if they had relevance to the network (based on either 0 or 1 within the matrix) they could be added.
2. How might you run a computer experiment to examine the deletion of a gene?
• All of the genes are listed within the spreadsheet. To examine the deletion of a gene, the spreadsheet would have to be manipulated. An easy way to do this would involve deleting the gene from the spreadsheet and observing its impact on the rest of the model.
3. Return to the chemostat. In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments?
• The least squares technique normalizes the data for a better fit. Looking at the ter Schure et al papers, almost all the figures exhibit a linear trend either up or down. Because of this, a linear best-fit line could help when attempting to approximate parameters.
4. Read over the Vu and Vhoradsky paper in this week's reading list. Is the method of the paper different from that in class? If so, in what ways?
• Vu and Vhoradsky utilized a similar method to our own, however, they are working with a larger gene pool. Their modeling is not much different than ours, aside from the fact that they used polynomial fit. Part of their reason for using this was because the data they used was only available as averages, whereas we initially began with raw data before scaling it.

Anthony J. Wavrin Week 13

1. How might you change the network to add more genes/transcription factors?
• Using the adjacency matrix in the excel sheet, you would add more genes/transcription factors on both the row and the columns to add both what transcription factors it regulates and what other transcription factors would regulate it using a binary code.
2. How might you run a computer experiment to examine the deletion of a gene?
• To examine the deletion of a gene, you would want to put zeros in the column of the gene to indicate that it does not affect any transcription factors which would mimic the deletion of the gene.
3. In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments?
• Least squares technique is used to minimize the variation between the line of best fit and the experimental data. Thus, using the same technique used in class today, one could derive the equations that were used in that model thousands of times changing the variables until the lowest least square value is obtained. We could then find the “correct” parameters for the model of the ter Schure et al experiment.
4. Is the method of the paper different from that in class? If so, in what ways?
• It is a little difficult to go through the paper and compare it to our method since it seems we have not totally finished our model. However, I found that they actually made a model for the degradation portion of the model while we used know values of the protein degradation at room temperature. Additionally, they used polynomial fit (regression) to the test the relationships as well as the least squares technique, we only used the latter. They also approximated regulated gene profiles by a polynomial of a degree n, which to my knowledge we did not do. Lastly, outside of the method of modeling, they were focusing on the cell cycle as opposed to cold shock.

Anthony J. Wavrin 22:38, 18 April 2013 (EDT)

Paul Magnano Week 13

1. How might you change the network to add more genes/transcription factors?
• You could add more rows and columns to the excel spreadsheet to account for the additional genes/transcription factors that you wanted to add.
2. How might you run a computer experiment to examine the deletion of a gene?
• If we were to use matlab like we did today to produce our graphs, we could either replace the 1 with a 0 in the excel sheet for a particular gene (represents deletion of the gene) or we could simply delete that genes row/column and then run the matlab script to get results to examine the genes deletion.
3. In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments?
• Since the concept behind the least squares technique relies on approximation, you could insert a best fit line into the ter Schure data to approximate the parameters of the ter Schure experiment.
4. Is the method of the paper different from that in class? If so, in what ways?
• The method an analysis that the paper used was similar to ours, with the exception that they made a model for the degredation portion while we used known values of the protein degradation. The fact they used polynomial fit also differed from our model, as well as the larger size of their gene pool.

Paul Magnano 23:58, 18 April 2013 (EDT)

Kasey E. O'Connor Week 13

1. How might you change the network to add more genes/transcription factors?
• To add more genes and transcription factors, you would just have to add more rows and columns to the matrix on Excel. For each added factor, you would have to figure out what it regulates and what regulates it. You could use the results from YEASTRACT to see how they interact with one another, especially since these are already in terms of 0s and ls.
2. How might you run a computer experiment to examine the deletion of a gene?
• To account for the deletion of the gene, you could either put 0's in every row and column associated with the gene, which shows that it does not affect or is not affected by another gene, or just completely delete the gene from the spreadsheets.
3. In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments?
• Because the data from the ter Schure paper was relatively linear, you could insert a best fit line into the data from the experiment. Then the least squares technique could be run to the equation of the line until it is closest to the data gathered by ter Schure. This would then result in the most accurate parameters we could get to model the data without knowing the exact parameters in the paper.
4. Read over the Vu and Vhoradsky paper in this week's reading list. Is the method of the paper different from that in class? If so, in what ways?
• The Vu and Vhoradsky paper looked at a larger amount of genes for their data, which was different than the 23 we analyzed in our method. Their method was very similar to ours, however they used a polynomial fit in addition to the least squares technique, while we only used the least squares.

Kasey E. O'Connor 23:59, 18 April 2013 (EDT)

Elizabeth Polidan Week 13

1. How might you change the network to add more genes/transcription factors?
• We have already prepared genes/transcription factors to add to this network. That was one of the results of last week’s assignments. They now need to be added to the input spreadsheet, along with parameter estimates. Some of these parameter estimates, such as the degradation rates, must be obtained from the literature. Some of the estimates, such as network weights, are guesses. Then the Matlab script can be rerun.
2. How might you run a computer experiment to examine the deletion of a gene?
• Just as you add genes to the network, you can subtract genes. Create an experimental copy of all of the scripts and spreadsheet, and remove the data for a certain gene. You can then run the script and see how the change propagates throughout the results.
3. In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments?
• You can use the least squares technique to better determine the expression levels of GAP1, PUT4, GDH1, GDH2, GLN1, HIS4, and ILV5 from the ter Schure paper.
4. Read over the Vu and Vhoradsky paper in this week's reading list. Is the method of the paper different from that in class? If so, in what ways?
• The V&V paper used a similar approach, including solving their differential equation using ode45 in Matlab. There were a few differences. They had many more data (time) points to fit, and used a 6-degree polynomial to approximate the expression of the genes. (6! ack! That is an extremely complex polynomial on such a relatively small number of data points.)

Elizabeth Polidan 01:33, 19 April 2013 (EDT)

Helena M. Olivieri| Helena M. Olivieri Week 13 Journal

1. How might you change the network to add more genes/transcription factors?
• Because last week we assessed which unique genes we would like to add to YEASTRACT model, it will soon be necessary to add actually add our genes to our Excel matrix. It will be necessary to, thus, continue to use YEASTRACT to generate a matrix that will make connections between the new unique genes and the previously studied genes. Results from already produced matrices may be inserted into our current Excel file.
2. How might you run a computer experiment to examine the deletion of a gene?
• Removing the gene completely or manipulating the amount of gene regulation in the excel file could emulate a gene deletion.
3. In concept, without worrying about creating matlab code, how might you use the least squares technique to get parameters from the ter Schure et al experiments?
• Least squares is a standard approach of approximating a solution. Data fitting can be identified as the most important application in terms of the least squares technique. Thus, because the ter Schure experiments tend to present linear data it would be appropriate to determine the parameters by using a linear best fit trend line.
4. Read over the Vu and Vhoradsky paper in this week's reading list. Is the method of the paper different from that in class? If so, in what ways?
• There were really only two differences I was able to note between the method of the paper and that of our class. The differences I saw were the use of a nonlinear model and a greater quantity of studied genes.