# Desireegonzalez Week 7

### Purpose

• The purpose of this weeks assignment was to learn how to interpret and analyze the data and figures given on the output files from the MATLAB run of our model during Week 6. This weeks assignment focuses on looking at the data with the assistance of GRNight to be able to visualize the weighted network graphs and then be able to propose a change in future experiments to keep studying the genes and their expression further.

### Methods

1. For model running methods using MATLAB, refer to the methods from the Week 6 Assignment Page.
2. To determine the LSE of the model data, the value was looked up on the "optimization_diagnostics" worksheet of the output workbook.
3. On this "optimization_diagnostics" worksheet the minLSE value was also looked up.
4. Then the LSE:minLSE ratio, was calculated by dividing the LSE by the minLSE; these ratio values will then be compared for the whole class.
5. Next, bar charts of the b and P parameters were created using Microsoft Excel.
6. Lastly, the output Excel file from Week 6, was uploaded onto GRNsight.
• Using the dropdown menu on the left the option to display data on the nodes (boxes) was chosen.
• The actual data for a strain was compared with the simulated data from the same strain by looking at the heatmap. If the model fits the data well, the color heatmap superimposed on the node will match top and bottom. If the fit is less good, the colors will not match.
• The individual expression plots were then compared to the heatmap data to determine if the line that represents the simulated model data was similar in its result of being a good or bad fit to the individual data points.

### Results

#### Analyzing Results of First Model Run

1. What is the overall least squares error (LSE) for your model?
• LSE= 0.981743
• minLSE= 0.696258
• LSE:minLSE= 0.981743:0.696258 which is about 1.41002729.
2. You need to look at the individual fits for each of the genes in your model. Which genes are modeled well? Which genes are not modeled well?
• When looking at the wt_log2_expression data versus the wt_log2_optimized_expression data the genes were depicted to model well when they had similar color on the heatmap on GRNsight; different colors on the heatmap depicted that the genes did not model well.
• AFT2: The model did not graph too well, since the optimized value resulted in the same color of red but in a much darker hue.
• ERT1: The model graphed well; both were similar shades of red.
• FHL1: The model did not graph well since the optimized values resulted in a blue color that was different from the red of the initial guess.
• GAL3: The model did not graph well since the optimized values resulted in a red color that was different from the blue of the initial guess.
• GCN4:The model did not graph well since the optimized values resulted in a blue color that was different from the red of the initial guess.
• GLN3: The model graphed well; both were similar shades of red.
• HAP4: The model did not graph well since the optimized values resulted in a red color that was different from the blue of the initial guess.
• IFH1: The model did not graph too well, since the optimized value resulted in the same color of red but in a much lighter hue than the initial guess.
• MBP1: The model graphed well; both were similar shades of red.
• PHO2: The model graphed well; both were similar shades of blue.
• SUM1: The model graphed well; both were similar shades of red.
• SUT2: The model graphed well; both were similar shades of red.
• TOD6: The model did not graph too well, since the optimized value resulted in the same color of red but in a darker hue.
• YGR067C: The model graphed well; both were similar shades of red.
• ZAP1: The model did not graph too well, since the optimized value resulted in the same color of red but in a much lighter hue than the initial guess.
• When looking at the dgln3_log2_expression data versus the dgln3_log2_optimized_expression data the genes were depicted to model well when they had similar color on the heatmap on GRNsight; different colors on the heatmap depicted that the genes did not model well.
• AFT2: The model graphed well; both were similar shades of red.
• ERT1: The model graphed well; both were similar shades of red.
• FHL1: The model did not graph well since the optimized values resulted in a blue color that was different from the initial guess that had some red.
• GAL3: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• GCN4: The model did not graph well since the optimized values resulted in a blue color that was lighter than the initial guess that dark blue.
• GLN3: The model graphed well; both were similar shades of red.
• HAP4: The model did not graph well since the optimized values resulted in a red color that was lighter than the initial red guess.
• IFH1: The model did not graph well since the optimized values resulted in a red color that was darker and different from the initial guess that had some blue.
• MBP1: The model graphed well; both were similar shades of red.
• PHO2: The model graphed well; both were similar shades of blue.
• SUM1: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• SUT2: The model graphed well; both were similar shades of red.
• TOD6: The model did not graph too well, since the optimized value resulted in the same color of red but in a darker hue than the initial guess.
• YGR067C: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• ZAP1: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• When looking at the dhap4_log2_expression data versus the dhap4_log2_optimized_expression data the genes were depicted to model well when they had similar color on the heatmap on GRNsight; different colors on the heatmap depicted that the genes did not model well.
• AFT2: The model did not graph well since the optimized values resulted in a red color that was lighter from initial guess that had dark red.
• ERT1: The model graphed well; both were similar shades of red.
• FHL1: The model graphed well; both were similar shades of blue.
• GAL3: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• GCN4: The model did not graph well since the optimized values resulted in a blue color that was lighter and different from the initial guess that had some red.
• GLN3: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• HAP4: The model did not graph well since the optimized values resulted in a red color that was different from the initial guess that blue.
• IFH1: The model graphed well; both were similar shades of red.
• MBP1: The model graphed well; both were similar shades of red.
• PHO2: The model graphed well; both were similar shades of blue.
• SUM1: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• SUT2: The model did not graph too well, since the optimized value resulted in the same color of red but in a lighter hue than the initial guess.
• TOD6: The model did not graph too well, since the optimized value resulted in the same color of red but in a darker hue than the initial guess.
• YGR067C: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• ZAP1: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• When looking at the dzap1_log2_expression data versus the dzap1_log2_optimized_expression data the genes were depicted to model well when they had similar color on the heatmap on GRNsight; different colors on the heatmap depicted that the genes did not model well.
• AFT2: The model did not graph well since the optimized values resulted in a red color that was a different shade from the one depicted in the initial guess.
• ERT1: The model graphed well; both were similar shades of red.
• FHL1: The model did not graph well since the optimized values resulted in a blue color that was lighter from initial guess that had some red.
• GAL3: The model did not graph well since the optimized values resulted in a blue color that was lighter from initial guess that had dark blue and red.
• GCN4: The model did not graph well since the optimized values resulted in a blue color that was lighter from initial guess that had some red.
• GLN3: The model graphed well; both were similar shades of red.
• HAP4: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• IFH1: The model did not graph well since the optimized values resulted in a red color that was darker from initial guess that had light red and blue.
• MBP1: The model did not graph well since the optimized values resulted in a red color that was lighter from initial guess that had some dark red.
• PHO2: The model graphed well; both were similar shades of blue.
• SUM1: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
• SUT2: The model did not graph well since the optimized values resulted in a red color that was darker from initial guess that had dark red.
• TOD6: The model did not graph well since the optimized values resulted in a red color that was lighter from initial guess that had dark red.
• YGR067C: The model graphed well; both were similar shades of blue.
• ZAP1: The model did not graph well since the optimized values resulted in a red color that was lighter and different from the initial guess that had some blue.
3. Upload your output Excel spreadsheet to GRNsight. Use the dropdown menu on the left to choose the data you will display on the nodes (boxes). Compare the actual data for a strain with the simulated data from the same strain. If the model fits the data well, the color heatmap superimposed on the node will match top and bottom. If the fit is less good, the colors will not match.
4. How many arrows are incoming to the node?
• AFT2:1 Repression arrow
• ERT1: 1 Self-Activation arrow
• FHL1: 0
• GAL3: 2 Repression arrows
• GCN4: 1 Activation arrow
• GLN3: 0
• HAP4: 0
• IFH1: 0
• MBP1: 1 Repression arrow
• PHO2: 2 Repression arrows, 1 Grey arrow, 1 Activation Arrow
• SUM1: 1 Repression arrow
• SUT2: 1 Repression arrow
• TOD6:2 Repression arrows
• YGR067C: 1 Grey arrow
• ZAP1: 1 Activation Arrow
5. What is the ANOVA Benjamini & Hochberg corrected p value for the gene?
• See attached files for table with ANOVA B-H p values.
6. Is the gene changing its expression a lot or is the log2 fold change mostly near zero?
• The genes that seemed to have the most change in expression are AFT2, TOD6, SUT2, SUM1, and MBP1; these genes have the largest difference in the heights of the production graphs for the initial guesses compared to the optimized data.
7. Make bar charts for the b and P parameters
• See attached files below for b and P bar charts.
8. Is there something about these parameters that explains the goodness of fit for the individual genes?
• There doesn't seem to be anything in these parameters that explains why a gene does or does not fit the model well.

### Future Experiments:Tweaking the Model and Analyzing the Results

• The next time the model is run in silico, I would like to edit the model by deleting the ZAP1 gene. With this deletion, we would hope to see better optimization values in a second run of the model. This deletion will hopefully create even more accurate values of p, b, and w for the gene network.

### Scientific Conclusion

• In conclusion, this weekly assignment's purpose was fulfilled since I was able to make an attempt in interpreting the data and figures given on the output files from the MATLAB run of our model during Week 6. The assignment's purpose was also fulfilled since a visualization of the weighted network was able to be created using GRNsight and we were able to obtain more precise values of p and b for the genes as depicted in the output files and images.

### Acknowledgments

• I communicated with my homework partners, Ava and Brianna through text message in order to figure out what portion of the model we would want to tweak for the second run during Week9; we wanted to be on the same page when editing the model to determine if the second run resulted in similar changes to the output data.
• I also communicated with Dr. Dahlquist and Dr. Fitzpatrick for assistance when I was confused on how to properly read and interpret the figure that we given as part of the output files from the Week 6 data.

Except for what is noted above, this individual journal entry was completed by me and not copied from another source. Desireegonzalez (talk) 22:01, 6 March 2019 (PST)