Natalie Williams Fall Electronic Notebook 2014
GRNmap Fall 2014
September 2014
September 18, 2014
Data Set Up
Openwetware familiarization: I became familiar with openwetware code and programing
MATLAB procedure: the MATLAB procedure that was written contains the instructions in using it to receive the output with the optimization network weights of the system.
Network
Ten random networks were made from the original network.
- The original network Excel file was used, and each cell on the network sheet had the following formula in it:
- =IF(RAND()<0.1134,1,0)
- This procedure was done ten times to get these ten random networks
- Each network was saved as rand# (1 - 10)
September 25, 2014
The random network Excel files were put into MATLAB to be run to get the optimized weights of the network
- The file is saved as the final name with _output.xls
Opening the file, the weights of these networks was found on the optimized_network_weights sheet.
Visualization of the Networks
These files had to be re-saved as .xlsx in order to upload them to GRNSight. GRNSight visualizes the networks and because there are varying numbers that suggest how much one gene controls another, the resulting output has different colors. After each individual random network was visualized, it was compared to the original network. For better analysis, the same order of the proteins was used to see the different connectivities.
October 2014
October 2, 2014
Information that could have been gathered from comparing the Original network to the 10 Random ones was found. This information includes:
Nodes: the positive and the negative
Frequencies: the In and Out degrees. These were how often one gene controlled other genes. It was found through the following equations:
- =COUNTIF(B2:B22,"<>"&0)
- =COUNTIF(B3:V3,"<>"&0)
- From this, the frequencies were found by looking at how often 0 appeared or 1 appeared, etc.
- For example, =COUNTIF(B23:V23,"=0") for In Degree to see how often 0 occurred
Next, bar graphs were used to compare the weighted networks between a random network and the original network.
After that, the minimum and maximum values from each random network was found.
- The minimum was found using: =MIN(B2:V22)
- The maximum was found using: =MAX(B2:V22)
The sum was found of the entire worksheet of the optimized_network_weights.
- =SUM(B2:V22)
The average of the worksheet was also taken for the entire matrix.
- =AVERAGE(B2:V22)
We used this information to see if there were any key factors to what made the original network the one that we accepted. We hoped that it would shed some light on what key differences were between the random networks and the original one.
October 9, 2014
I was out of town, so there was nothing I needed to do specifically for this week.
October 16, 2014
I began the process of the forward simulations of the networks. I had to isolate the deletion strains and see if there was any resemblance between the wild type strain with the four deletion strains.
October 23, 2014
All the bugs in the system were noted and written down to be fixed.
The forward simulations were rerun. The production and degradation rates from the output were inserted into each of the individual strains. For the network weights of the individual strains, the output from the general workbook sheet, optimized_network_weights, was used.
The deletion strains needed hard 0's across their row on their worksheet. On the optimization_parameters sheet, the following things needed to be altered:
- iestimate = 0.00+E0
- fixed_b = 0
- strains:
- wt/3/0
- dcin5/4/3, where the first number is the sheet, and the second number is the row of the gene within the sheet
- This controlled which strains would be shown after the workbook was run through MATLAB
Network_b sheet used the optimized_network_b from the general workbook output was used for each individual strain.
October 30, 2014
The Real WT individual strain was compared to the forward simulation WT and deletion strains.
I made a list of transcription factors of the individual strains that did not compare well with the real WT individual. Those transcription factors were going to be looked at more closely and might have been taken away. The parts that I compared were the data points and the fit of the line.
November 2014
November 6, 2014
16 transcription factors were taken and run through YEASTRACT. However, the results have to be formatted in a way so that GRNSIght can visualize it the network that results.
The network I used was created with the following transcription factors: ARG80 CIN5 GLN3 HAP4 HMO1 NRG2 RSF2 RTG3 STB4 SWI4 TBF1 TOS8 TYE7 YHP1 YOX1 ZAP1
- Navigate to Generate Regulation Matrix [[1]] on the YEASTRACT
- Select the appropriate check boxes for the filters.
- Paste the list of transcription factors into the appropriate field.
- Paste a list of targets into the Target ORF/Genes field, or check the box to consider all ORF/Genes.
- Click the Generate button.
- In the results window that appears, click the link to download the Regulation matrix results file as a Semicolon Separated Values (CSV) file.
- Once you have downloaded the file, launch Microsoft Excel.
- Select the menu item, File > Open and select the file that you downloaded.
- Select Column A.
- Select the menu item, Data > Text to Columns...
- In the first window of the wizard that appears, select the radio button for "Delimited" and click Next.
- In the second window of the wizard that appears, check the box for "Other" under "Delimiters" and type a semicolon in the field to the right and click Finish.
- Select the menu item, File > Save As. Save the file as an Excel Workbook (.xlsx).
- The orientation of the matrix has to be flipped. A new worksheet must be created by clicking on the new worksheet icon at the bottom of the screen. Name this new worksheet "network".
- Select the adjacency matrix from the first worksheet and copy it to the clipboard. Go to the "network" worksheet and click on cell A1. Select the menu item Edit > Paste Special. In the window that appears, check the box "Transpose" and click OK.
- The labels for the genes in the columns and rows needs to match. The "p" of the gene names in the columns must be deleted.
- Paste the following text into cell A1 "rows genes affected/cols genes controlling".
- Save your work, which is now ready for loading into GRNsight. The original sheet can be deleted if you want.
Results
GRNSight v1.8 had to be used to visualize the networks. Only four of the input selection choices gave network connections among the listed transcription factors.
Documented DNA Binding Evidence
- 15 genes
- 58 edges
Documented DNA Expression
- 15 genes
- 31 edges
DNA Expression plus Binding
- 15 genes
- 58 edges
DNA Expression and Binding
- 15 genes
- 4 edges
Potential with Motifs
- 0 genes
- 0 edges
Potential without Motifs
- 0 genes
- 0 edges
Documented plus Potential
- 0 genes
- 0 edges
Documented and Potential
- 0 genes
- 0 edges
November 13, 2014
I reran MATLAB to see if I got the same results as Dr. Fitzpatrick. I received the same results as Dr. Fitzpatrick. When each deletion strain was compared to the WT strain, the targeted genes that were supposed to be affected were.
Other Links
Back to User:Natalie Williams
To visit the Dahlquist Lab: click here