Falghane Week 7

Purpose

Analyzing the results input results of models obtained from experiment done from week 4 to week 6 in order to understand the gene networking better.

Methods

• squares error (LSE) was found from "optimization_diagnostics" worksheet of output workbook..
• The overall individual fits for each of the genes in the model and weather the gene was well modeled or not was found by analyzing the graphs and using GRNsight. The model was considered to fit the data well if the color heat map superimposed on the node matched top and bottom. If the fit was less good, the colors did not match.

"Tweak"-Deletion of gene CUP2

• CUP2 was deleted from all the worksheets in the output excel file [1].
• After that, all the outputs previous results were deleted.
• The file was then renamed to "input workbook_FA_CUP2deletion."
• It was then uploaded to GRNsight and gave the resulting new network between genes.
• After that the model was ran through MATLAB in order to give the results of deleting CUP2 and the effect it has in other genes.

Results

Analyzing Results of First Model Run

• What is the overall least squares error (LSE) for your model? 1.0236
  • LSE:minLSE:1.0236/0.667 = 1.53
• 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?
  • Gene SUT1: it was modeled well as it is inclusive for the average for wt data, dgln3 data, and dhap4 data. Also from GRNsight the color of the model is very similar to that of the actual data.
  • Gene RGM1: Based on the graph, RGM1 model was a good representation. However, it was not a good representative of dhap4 and wt data. However, from the GRNsight the color heatmap superimposed on the node will match top and bottom which means that the model fit the data well.
  • Gene PDR1: from the graph, PDR1 was well modeled, also based on the GRnsight the color on the top of the nod is very similar to the one at the bottom which means that the model fits the data well.
  • Gene HOT1: HOT1 is well modeled proven by both the graph and color of nod in GRNsight.
  • Gene MGA2: MGA2 is well modeled. It has two different graphs because of the different distribution of genes. However, in GRNsight the nod color for MGA2 is the same for both actual and expected expression, therefore, it's a good model.
  • Gene MSN4: MSN4 is a good model, in GRNsight the color of the model is very similar to that of the actual data. Also from the graph, it fits in between wt data, dgln3 data, and dhap4 data.
  • Gene MSN2: It's a good model as shown in the graph and also by the GRNsight as the color heatmap superimposed on the node match top and bottom.
  • Gene MIG1: It's not a good model as shown by GRNsight as the top and bottom colors of the nod do not match.
  • Gene HAP1: It's a good model as shown in the graph and also by the GRNsight as the color heatmap superimposed on the node match top and bottom.
  • Gene GLN3: It's a good model as shown in the graph and also by the GRNsight as the color heatmap superimposed on the node match top and bottom.
  • Gene HAP4: It's not a good model as shown by GRNsight as the top and bottom colors of the nod do not match.
  • Gene AFT1: It's a good model as shown in the graph and also by the GRNsight as the color heatmap superimposed on the node match top and bottom.
  • Gene GCN4: It's not a good model as shown by GRNsight as the top and bottom colors of the nod do not match.
  • Gene GIS1: It's a good model as shown in the graph and also by the GRNsight as the color heatmap superimposed on the node match top and bottom.
  • Gene AFT2: It's a good enough model as shown in the graph and also by the GRNsight as the color heatmap superimposed on the node mostly match top and bottom.
  • Gene ADR1: It's a good model as shown in the graph and also by the GRNsight as the color heatmap superimposed on the node match top and bottom.
  • Gene CUP2: It's a good model as shown in the graph and also by the GRNsight as the color heatmap superimposed on the node match top and bottom.

Change in expression

• 9 of the 14 genes had big change in expression and they were the following:
  • SUT1, RGM1, MGA2, MSN4, MSN2, GLN3, HAP4, AFT1, GLN4, GLS1, AFT2.
  • 2 out of the 9 gene models that had a big change in expression were not well modeled.
  • 1 out of 6 gene models with no big change in expression were not well modeled.

Deletion of gene CUP2 Results

• The results from the deletion of the gene CUP2 could be find in the following link [2].

PowerPoint Presentation

• The PowerPoint Presentation could be found here [3]

Scientific Conclusion

based on the results, there is no relationship between change in expression and weather a gene is modeled well or not.

Acknowledgment

• I worked with Dr. Kam D. Dahlquist in class in analyzing the data.
  • Except for what is noted above, this individual journal entry was completed by me and not copied from another source.

Falghane (talk) 16:20, 6 March 2019 (PST)

References

• Instructions from Week 7 were followed .
• The following [4] contained the data that was analyzed.