Bklein7 Week 8

From LMU BioDB 2015
Jump to: navigation, search

Files Generated in the This Week's Analysis

Links to files below can all be found within the electronic notebook at the point where they were created. For easy access, they are listed here as well.

  1. Analyzed microarray data: File:Merrell Compiled Raw Data Vibrio BK 20151015.xls.
  2. For GenMAPP text file: File:Merrell Compiled Raw Data Vibrio BK 20151015- Tab Delimited.txt.
  3. Exceptions file: File:Merrell Compiled Raw Data Vibrio BK 20151015.EX.txt.
  4. Expression Dataset files:
  5. GO term MAPP: File:Polyribonucleotide nucleotidyltransferase activity.mapp.
  6. MAPPFinder Results:
  7. .gmf file: File:Merrell Compiled Raw Data Vibrio BK 20151025.gmf.

Statistical Analysis of Vibrio cholerae Microarray Data (Part 1)

Normalizing the Log Ratios for the Set of Slides in the Experiment

This section dictates the steps necessary to scale and center the raw microarray data:

  • To begin, I created a new Worksheet in my Excel file entitled "scaled_centered".
  • I went back to the original "compiled_raw_data" worksheet and copied over all the data into the new "scaled_centered" worksheet.
  • I inserted two rows in between the top row of headers and the first data row entitled "Average" (cell A2) "StdDev" (cell A3).
    • I computed the Average log ratio for each chip by inputting the following equation into cell B2 and then pasting it into the rest of the "Average" column: =AVERAGE(B4:B5224).
    • I computed the Standard Deviation of the log ratios on each chip by inputting the following equation into cell B3 and then pasting it into the rest of the "StdDev" column: =STDEV(B4:B5224).
  • I created a new set of headings for the scaled and centered data by copying over the data column headings and then pasting them to the right of the last data column. I edited the names of the columns so that they now read: A1_scaled_centered, A2_scaled_centered, etc.
  • In cell N4 (column with the heading A1_scaled_centered), I typed the following equation: <code)=(B4-B$2)/B$3</code>. In this case, we want the data in cell B4 to have the average subtracted from it (cell B2) and be divided by the standard deviation (cell B3). We use the dollar sign symbols in front of the "2" and "3" to tell Excel to always reference that row in the equation, even though we will paste it for the entire column of 5221 genes. Why is this important?
    • Adding the dollar sign before the 2 and 3 ensured that the equation for each individual gene in column B drew from the overall average and standard deviation for the column. Maintaining these overall values in the equation is critical to yielding scaled and centered outputs for each gene. If the dollar signs were not included, Excel would assume that for each gene, it would subtract by the value two rows above and then divide by the value one row above (e.g. for B80, the equation would be (B80-B78)/B79 when in reality we would want (B80-B2)/B3). This would yield extraneous results.
  • I copied this equation to the rest of the column and then adapted it for all "_scaled_centered" columns.

Performing Statistical Analysis on the Ratios

This section details the steps necessary to perform statistical analysis on the scaled and centered data produced in the section above:

  • For this section, I created a new worksheet and name it "statistics".
  • I went back to the "scaled_centered" worksheet and copied over both the first column ("ID") and copied the columns that are designated "_scaled_centered".
  • I deleted rows 2 and 3 where it says "Average" and "StDev" so that the data rows with gene IDs were immediately below the header row 1.
  • I created three new columns to the right of the copied ones with the following headers: "Avg_LogFC_A", "Avg_LogFC_B", and "Avg_LogFC_C".
  • I computed the average log fold change for the replicates for each patient by typing the equation:
=AVERAGE(B2:E2)
into cell N2 and then copying/pasting it into the rest of the three columns, adapting it as necessary.
  • I created a new column to compute the average of the averages with the header "Avg_LogFC_all". I then created the equation to compute the average of the three previous averages you calculated and paste it into this entire column.
  • I created a new column with the header "Tstat" to compute the T statistic for each scaled and centered average log ratio. In this column, I entered the following equation and pasted it into the rest of the column: 
=AVERAGE(N2:P2)/(STDEV(N2:P2)/SQRT(3))
  • I created one last column with the heading "Pvalue". In the cell below the label, I entered the following equation and then copied it into the rest of the column: 
=TDIST(ABS(R2),2,2)

Calculating the Bonferroni p value Correction

It is necessary to perform adjustments to the p value to correct for the multiple testing problem. Therefore, I first calculated the Bonferroni p value correction on the p values calculated in the above section:

  • I labelled the next two columns to the right of "Pvalue" with the same label: "Bonferroni_Pvalue".
  • I typed the equation =S2*5221 into the first Bonferroni column and pasted it throughout.
  • Next, I replaced any corrected p value that was greater than 1 with the number 1 by typing the following formula into the first cell below the second Bonferroni_Pvalue header: =IF(T2>1,1,T2).

Calculating the Benjamini & Hochberg p value Correction

The second p value correction I performed was the Benjamini & Hochberg correction, the methods of which are presented below:

  • I created a new worksheet named "B-H_Pvalue".
  • In this worksheet, I copied over the "ID" column from the previous worksheet into the first column of the new worksheet.
  • Next, I added a new column on the very left and named it "MasterIndex". Under this heading, I added a numbered list from 1 to 5221 (the number of genes on the microarray).
  • I copied over the unadjusted p values from your previous worksheet and pasted them into Column C.
  • I selected all of columns A, B, and C and sorted by ascending values on Column C.
    • This was done by clicking the sort button from A to Z on the toolbar and sorting by column C, smallest to largest.
  • I created a new column and labelled it with the header "Rank" in cell D1 to house another numbered list from 1 to 5221. Because this was done for the sorted columns, these values represented the p value ranks, smallest to largest.
  • I created a two new columns to calculate the Benjamini and Hochberg p value correction and labeled them "B-H_Pvalue" (cell E1 and F1).
    • In the first column, I inputted the following formula and copied it throughout the column: =(C2*5221)/D2.
    • In the second column, I inputted the following formula and copied it throughout: =IF(E2>1,1,E2).
  • Next, I selected columns A through F and sorted them by the MasterIndex in Column A in ascending order.
  • Finally, I copied column F into the next column on the right of your "statistics" sheet.

Preparing the File for GenMAPP

  • I inserted a new worksheet and named it "forGenMAPP".
  • I copied over the entirety of the "statistics" worksheet to this new worksheet.
  • I selected Columns B through Q (all the fold changes) and formatted the cells to show only 2 decimal places.
  • I selected all the columns containing p values and formatted the cells to show only 4 decimal places.
  • I deleted the left-most Bonferroni p value column.
  • I inserted a column to the right of the "ID" column entitled "SystemCode" and filled the entire column with the letter "N".
  • After making the above changes, I saved the file as "Text (Tab-delimited) (*.txt)".

Sanity Check: Number of Genes Significantly Changed

To verify the results of the data analysis performed on the Vibrio cholerae DNA microarray data, I assessed the number of genes that were significantly changed at various p value cut-offs and also compared my data analysis with the published results of Merrell et al. (2002). These analyses were done in the "forGenMapp" tab of the Excel spreadsheet used for the data analysis.

  • Assessing the number of genes significantly changed
    • To answer the questions below, I used custom filter to display only the "Pvalue" data that met specific criterion (e.g. < 0.05).
    • How many genes have p value < 0.05? and what is the percentage (out of 5221)?
      • 948 of the 5221 genes (18.2%) had p values that were less than 0.05
    • What about p < 0.01? and what is the percentage (out of 5221)?
      • 235 of the 5221 genes (4.50%) had p values that were less than 0.01
    • What about p < 0.001? and what is the percentage (out of 5221)?
      • 24 of the 5221 genes (0.46%) had p values that were less than 0.001
    • What about p < 0.0001? and what is the percentage (out of 5221)?
      • 2 of the 5221 (0.038%) genes had p values that were less than 0.0001
    • To apply a more stringent criterion to the p values, I performed the Bonferroni and Benjamini and Hochberg corrections to the unadjusted p values. I then filtered the columns "Bonferroni_Pvalue" and "B-H_Pvalue" columns to only display data that met specific criterion.
    • How many genes are p < 0.05 for the Bonferroni-corrected p value? and what is the percentage (out of 5221)?
      • No genes (0.00%) had p values less than 0.05 for the Bonferroni-corrected p value
    • How many genes are p < 0.05 for the Benjamini and Hochberg-corrected p value? and what is the percentage (out of 5221)?
      • No genes (0.00%) had p values less than 0.05 for the Benjamini and Hochberg-corrected p value
  • Assessing the number of genes with expression changes
    • The "Avg_LogFC_all" indicated the size of the gene expression changes and their direction. Positive values increased relative to the control, and negative values decreased relative to the control.
      • Keeping the (unadjusted) "Pvalue" filter at p < 0.05, I filtered the "Avg_LogFC_all" column to show all genes with an average log fold change greater than zero. How many are there? (and %)
        • 352 of the 5221 genes (6.74%) exhibited significant (p < 0.05) increases in gene expression
      • Keeping the (unadjusted) "Pvalue" filter at p < 0.05, I filtered the "Avg_LogFC_all" column to show all genes with an average log fold change less than zero. How many are there? (and %)
        • 596 of the 5221 genes (11.4%) exhibited significant (p < 0.05) decreases in gene expression
      • What about an average log fold change of > 0.25 and p < 0.05? (and %)
        • 339 of the 5221 genes (6.49%) exhibited significant (p < 0.05) increases in gene expression that exceeded an average log fold change of 0.25
      • Or an average log fold change of < -0.25 and p < 0.05? (and %)
        • 579 of the 5221 genes (11.1%) exhibited significant (p < 0.05) decreases in gene expression that were less than an average log fold change of -0.25
  • What criteria did Merrell et al. (2002) use to determine a significant gene expression change? How does it compare to our method?
    • Merrel et al. identified statistically significant changes in gene expression by imputing normalized intensity ratios for expression into the program Statistical Analysis for Microarrays (SAM). Within the program, they ran a two-class SAM analysis using the strain grown in vitro (class I) and each individual stool sample (class II). This test isolated expression levels that were of a significantly different magnitude (at least a twofold change) across all patient samples as statistically significant expression changes. This method for significance differed from ours in two primary ways:
      1. It judged significant differences based on order of magnitude changes in the normalized intensity rations (twofold change) as opposed to running T tests and calculating p-values.
      2. It looked for consistent significant changes in expression across all patient samples instead of looking for significance in averaged log fold changes for all patients.
  • For the GenMAPP analysis below, I used the fold change cut-off of greater than 0.25 or less than -0.25 and the unadjusted p value cut off of p < 0.05 for the analysis.

Sanity Check: Compare Individual Genes with Known Data

Merrell et al. (2002) report that genes with IDs: VC0028, VC0941, VC0869, VC0051, VC0647, VC0468, VC2350, and VCA0583 were all significantly changed in their data. I reviewed the "forGenMAPP" Excel worksheet to compare these findings to the results of my personal data analysis.

  • What are the fold changes and p values (of these genes)? Are they significantly changed in the analysis?
    • VC0028 (2 entries)
      • Fold changes: 1.65, 1.27
      • P values: 0.0474, 0.0692
      • Significantly changed? The expression of the first gene with this ID significantly changed, but that of the second gene with this ID did not.
    • VC0941 (2 entires)
      • Fold changes: 0.09, -0.28
      • P values: 0.6759, 0.1636
      • Significantly changed? The expression of neither gene with this ID significantly changed.
    • VC0869 (5 entries)
      • Fold changes: 1.50, 1.59, 1.95, 2.20, 2.12
      • P values: 0.0174, 0.0463, 0.0227, 0.0020, 0.0200
      • Significantly changed? The expression of all genes with this ID significantly changed.
    • VC0051 (2 entries)
      • Fold changes: 1.92, 1.89
      • P values: 0.0139, 0.0160
      • Significantly changed? The expression of both genes with this ID significantly changed.
    • VC0647 (3 entries)
      • Fold changes: -1.11, -0.94, -1.05
      • P values: 0.0003, 0.0125, 0.0051
      • Significantly changed? The expression of all genes with this ID significantly changed.
    • VC0468 (1 entry)
      • Fold change: -0.17
      • P value: 0.3350
      • Significantly changed? The expression of this gene did not significantly change.
    • VC2350 (1 entry)
      • Fold changes: -2.40
      • P value: 0.0130
      • Significantly changed? The expression of this gene significantly changed.
    • VCA0583 (1 entry)
      • Fold change: 1.06
      • P value: 0.1011
      • Significantly changed? The expression of this gene did not significantly change.

MAPPFinder Analysis of Vibrio cholerae Microarray Data (Part 2)

Mapping Onto Biological Pathways (GenMAPP & MAPPFinder)

Before beginning the mapping process, it is necessary to load the correct gene database into GenMAPP:

  • I download the 2009 Vibrio cholerae gene database by following this link to the XMLPipeDB SourceForge Download page. My homework partner Veronica downloaded the 2010 version.
  • I saved the above file into the folder C:\GenMAPP 2 Data\Gene Databases and extracted it.
  • Within GenMAPP, I loaded the 2009 gene database by selecting Data > Choose Gene Database and choosing the appropriate file from the directory C:\GenMAPP 2 Data\Gene Databases.

GenMAPP Expression Dataset Manager Procedure

Once the appropriate gene database has been loaded into GenMAPP, the expression dataset can be uploaded and configured:

  • I opened the Expression Dataset Manger from the Data drop-down list in GenMAPP.
  • I selected New Dataset from the Expression Datasets menu and choose the tab-delimited text file formatted for GenMAPP (.txt).
  • Upon specifying that all data was numerical, the Expression Dataset Manager converted my data to .gex file. This process took approximately one minute to complete. In addition to converting the data to a .gex file, an exceptions file (.EX.txt) was also produced, as 772 errors were reportedly detected in the raw data.
    • The .gex file generated can be found here: File:Merrell Compiled Raw Data Vibrio BK 20151015.gex
    • The .EX file generated can be found here: File:Merrell Compiled Raw Data Vibrio BK 20151015.EX.txt
    • Record the number of errors. For your journal assignment, open the .EX.txt file and use the Data > Filter > Autofilter function to determine what the errors were for the rows that were not converted.
      • GenMAPP errors message.PNG
      • The above screenshot shows the error message I received after using the Expression Dataset Manager to convert my raw data file. A reported 772 errors were detected in my raw data.
      • Upon opening the .EX.txt file, I found the error message "Gene not found in OrderedLocusNames or any related system." repeated several times.
        • To determine if this was the only error message, I uploaded the exceptions file to my LMU CMSI directory using the command scp /Users/brandonklein/Desktop/Merrell_Compiled_Raw_Data_Vibrio_BK_20121015.EX.txt bklein7@my.cs.lmu.edu:~bklein7. In the command line, I used the following command sequence to determine how often the "Gene not found..." error message was repeated: grep "Gene not found"Merrell_Compiled_Raw_Data_Vibrio_BK_20121015.EX.txt | wc. This yielded the output 772 23932 145791, confirming that the above error message was responsible for all 772 reported errors.
      • It is likely that you will have a different number of errors than your partner who is using a different version of the Vibrio cholerae Gene Database. Which of you has more errors? Why do you think that is?
        • With the 2009 Vibrio cholerae gene database loaded into GenMAPP, I encountered 772 errors when converting the analyzed Merrell et al. microarray data. When converting the exact same data with the 2010 Vibrio cholera gene database loaded into GenMapp, Veronica encountered 121 errors. All error messages for both of us were the same: "Gene not found in OrderedLocusNames or any related system". Therefore, I had more errors. I believe this occured because the 2009 version of the Vibrio cholerae database was less developed/complete than the more recent 2010 version. Thus, it likely had less Gene listings, and therefore less of the Ordered Locus Names used by Merrell matched to the database.
  • I customized the new Expression Dataset by creating a Color Sets= with instructions to GenMAPP for displaying data on MAPPs. The new Color Set was entitled "LogFoldChange".
    • First, I created a criterion for this color set to label genes that demonstrated a significant increase in their expression.
      • I specified the Gene value as "Avg_LogFC_all" for the Vibrio dataset.
      • I activated the Criteria Builder by clicking the New button and named the criterion "Increased".
      • I selected the color for this criterion as red using the color box.
      • I stated the criterion as follows and added it to the Criteria List: [Avg_LogFC_all] > 0.25 AND [Pvalue] < 0.05
    • Second, I created a criterion for this color set to label genes that demonstrated a significant decrease in their expression.
      • I specified the Gene value as "Avg_LogFC_all" for the Vibrio dataset.
      • I activated the Criteria Builder by clicking the New button and named the criterion "Decreased".
      • I selected the color for this criterion as green using the color box.
      • I stated the criterion as follows and added it to the Criteria List: [Avg_LogFC_all] < -0.25 AND [Pvalue] < 0.05
  • Upon entering these color sets, I savedthe entire Expression Dataset by selecting Save from the Expression Dataset menu.

MAPPFinder Procedure

  • I launched the MAPPFinder program from within GenMAPP and ensured that the 2009 Gene Database was still loaded into GenMAPP.
  • I clicked on the button "Calculate New Results" followed by "Find File", at which point I chose my .gex file.
  • Veronica and I both chose to filter the data with the "Decreased" criterion present within the LogFoldChange Color Set.
  • I checked the boxes next to "Gene Ontology" and "p value", specified the results file, and then clicked "Run MAPPFinder".
    • This analysis took several minutes to complete.
  • I clicked on the menu item "Show Ranked List" to see a list of the most significant Gene Ontology terms.
    • List the top 10 Gene Ontology terms.
      • TOP10.PNG
      1. protein folding
      2. chorismate metabolic process
      3. aromatic amino acid family biosynthetic process
      4. unfolded protein binding
      5. cytoplasm
      6. intracellular part
      7. localization
      8. nucleotide catabolic process
      9. locomotion
      10. aromatic amino acid family metabolic process
    • Compare your list with your partner who used a different version of the Gene Database. Are your terms the same or different? Why do you think that is?
      • Veronica's top 10 Gene Ontology terms using the 2010 Vibrio database were as follows:
          1. signal transduction
          2. molecular transducer activity
          3. signal transducer activity
          4. membrane
          5. peptidyl-histidine modification
          6. peptidyl-histidine phosphorylation
          7. two-component sensor activity
          8. protein histidine kinase activity
          9. peptidyl-amino acid modification
          10. phosphotransferase activity, nitrogenous group as acceptor
      • The above list of Veronica's top 10 Gene Ontology terms was entirely different from my own, despite the fact that we both input the same microarray data into GenMAPP. However, there are several reasons that explain why this happened. First, Veronica used a more recent version of the Vibrio database (2010 vs. 2009), which matched GO terms to a large number of genes included in the microarray data that did not perviously have GO terms in the 2009 database. As evidence of this, I only retrieved GO terms for two of the gene ID's in the question below using the 2009 database, whereas Veronica retrieved GO terms for each of the gene ID's. Additionally, it is possible that new GO terms were added to the 2010 database that were not present prior. For example, when I searched one of the GO terms Veronica found associated with the gene ID VC0028, dihydroxy-acid dehydratase activity, I did not find any matches to this result. Finally, each time a MAPP is generated, a slightly different list of top Gene Ontology terms can be produced. I tested this by running multiple MAPPFinder analyses on my microarray data using the same 2009 database and retrieved slightly different results more than once.
  • I used MAPPFinder to find the Gene Ontology term(s) with which the following genes mentioned by Merrell et al. (2002) were associated with: VC0028, VC0941, VC0869, VC0051, VC0647, VC0468, VC2350, and VCA0583. This was done by typing the identifiers for these genes into the MAPPFinder browser gene ID search field, choosing "OrderedLocusNames" from the drop-down menu to the right of the search field, and clicking on the GeneID Search button. The GO term(s) that were associated with the genes were highlighted in blue. List the GO terms associated with each of those genes. (Note: they might not all be found.) Are they the same as your partner who is using a different Gene Database? Why or why not?
    • Associated GO terms
      • VC0028: "No MAPPs or GO terms could be found for that OrderedLocusNames ID."
      • VC0941: "No MAPPs or GO terms could be found for that OrderedLocusNames ID."
      • VC0869: "No MAPPs or GO terms could be found for that OrderedLocusNames ID."
      • VC0051: "No MAPPs or GO terms could be found for that OrderedLocusNames ID."
      • VC0647: mRNA catabolic process, RNA processing, cytoplasm, RNA binding, 3'-5'-exoribonuclease activity, transferase activity, nucleotidyltransferase activity, and polyribonucleotide nucleotidyltransferase activity.
      • VC0468: "No MAPPs or GO terms could be found for that OrderedLocusNames ID."
      • VC2350: "No MAPPs or GO terms could be found for that OrderedLocusNames ID."
      • VCA0583: transport, outer membrane-bounded periplasmic space, and transporter activity.
    • Comparison with Veronica's results using the 2010 Vibrio database
      • As mentioned in the answer to the previous question, Veronica matched all of the above GO terms to gene IDs in the 2010 database. Thus, the 2010 database must have been updated to include a more complete listing of GO terms associated with specific gene IDs, perhaps reflecting research advancements. In addition to this, some of the GO terms we retrieved for the same gene IDs (such as VCA0583) were not the same. This also suggests that more refined or experimentally proven GO terms were matched with genes previously associated with GO terms in the 2010 database.
  • VC0647 Investigation: List the name of the GO term you clicked on and whether the expression of the gene you were looking for changed significantly in the experiment.
    • I clicked on the GO term "polyribonucleotide nucleotidyltransferase activity". The MAPP this produced only contained the gene "PNP_VIBCH", which I found in the [UniProt Database]. The expression of this gene, referenced by the gene ID VC0647, significantly decreased in the microarray experiment.
    • I Double-clicked on the above gene box to find out more about the gene PNP_VIBCH. Click on the links to find out the function of this gene.
      • According to the [UniProt Database], the function of this gene is to degrade mRNA. It does this by coding for the production of the protein polyribonucleotide nucleotidyltransferase, which "catalyzes the phosphorolysis of single-stranded polyribonucleotides processively in the 3'- to 5'-direction".
    • The MAPP created while investigating polyribonucleotide nucleotidyltransferase activity can be found here: File:Polyribonucleotide nucleotidyltransferase activity.mapp.
  • Overall MAPPFinder results file: File:Merrell Compiled Raw Data Vibrio BK 20151015-Criterion1-GO.txt.
  • I opened a copy of the .txt file listed above in Excel to filter the results.
    • There were rows of information that gave background information on how MAPPFinder made the calculations at the top of the Excel file. Compare this information with your partner who used a different version of the Vibrio Gene Database. Which numbers are different? Why are they different?
      • When comparing the background information, I found several lines in my output file that were distinct from Veronica's using the 2010 Vibrio database:
        1. Veronica matched a dramatically larger number of probes (3159 vs. 578) to the decreased criterion. This was unexpected, as I had anticipated our results for this to be equal given that we input the same microarray data. However, it turned out there was a flaw in Veronica's decreased criterion, which read: [Avg_LogFC_All]>-0.25 AND [Pvalue]>0.05 instead of [Avg_LogFC_All] < -0.25 AND [Pvalue]<0.05.
        2. Due to the issue above, an equally dramatic difference was present between the number of probes meeting the criterion that matched with a UniProt ID. This was expected, as the evidence during this exercise suggested that the 2010 database was more complete, including more gene IDs. Indeed, even with the criterion error, the percentage of probes that met her criterion and also matched with a UniProt ID (97.37%) was greater than mine (81.83%). This supports the conclusion that the 2010 database included more gene IDs.
      1. The number of probes meeting the criterion AND linking to a GO term was different for both of us in much the same was as the example above. Although the strict numerical difference was exaggerated due to error (1571 vs. 254), the percentages still supported the fact that the 2010 database matched gene IDs to GO terms more frequently (49.73% vs. 43.94%).
      2. Finally, both the number of probes in the entire dataset that linked to UniProt IDs as well as to GO terms followed the trends already detailed above, proving those points. The 2009 database I used did in fact include less links to UniProt IDs and GO terms on the whole than the 2010 database.
    • I used the following filters to show the top 20 GO terms represented in my data for both the "Increased" and "Decreased" criteria:
Z Score (in column N) greater than 2
PermuteP (in column O) less than 0.05
Number Changed (in column I) greater than or equal to 5 AND less than 100
Percent Changed (in column L) greater than or equal to 30%
  • Are any of your filtered GO terms closely related to one another, meaning are they a direct child or parent to another term in the list? You can judge this by comparing your spreadsheet with the MAPPFinder browser. Highlight the terms that fit this relationship with the same color in your Excel spreadsheet.
  • Interpret your results. Look up the definitions for any GO terms that are unfamiliar to you. The "official" definitions for GO terms can be found at http://www.geneontology.org. You can use one of the online biological dictionaries as a supplement, if needed. Write a paragraph relating the results of this GO analysis to the experiment performed (comparing laboratory-grown and patient-derived Vibrio cholerae. You need to give a biological interpretation of what do each of these GO terms in your filtered list have to to with the pathogenecity of the bacterium? You may consult with your partner on this, but your explanation on your individual journal page needs to be in your own words. This is where the real "brain power" comes in with interpreting DNA microarray data. Even experienced scientists struggle with this part. Use your creativity as a scientist to stretch your brain in this question.
    • To begin the interpretation, I used the UniProt database in tandem with my MappFinder results and the filtered Excel file to discern both the function of specific genes included under the top 20 GO terms and whether those genes were down-regulated or up-regulated in the experiment. To facilitate this, I grouped the GO terms that were closely related according to the highlights I made in the filtered Excel file. For the sake of simplicity, I included just an interpretation of the results from this investigation for each group below.
      • Group 1: protein folding and unfolded protein binding
        • Decreased expression of genes that promote the proper folding of proteins in stress conditions.
          • Decreased synthesis of chaperone proteins.
        • Increased pepsidase activity, which promotes the hydrolysis of proteins.
      • Group 2: cis-trans isomerase activity and peptidyl-prolyl cis-trans isomerase activity
        • Decreased expression of genes that promote proper protein folding and prevent protein aggregation (similar to Group 1).
      • Group 3: zinc ion binding
        • Decreased synthesis of proteins involved in glycolytic processes.
        • Increased pepsidase activity.
      • Group 4: sugar:hydrogen symporter activity, solute:hydrogen symporter activity, and cation:sugar symporter activity
        • Decreased sugar-dependent active transport into membrane.
      • Group 5: protein-N(PI)-phosphohistidine-sugar phosphotransferase activity, sugar transmembrane transporter activity, and carbohydrate transmembrane transporter activity
        • Effects similar to Group 4.
      • Group 6: phosphoenolpyruvate-dependent sugar phosphotransferase system and carbohydrate transport
        • Decreased intake of sugars from the external environment.
        • Effects similar to Group 4.
      • Group 7: translation regulator activity and translation factor activity, nucleic acid binding
        • Decreased stimulation of aa-tRNA binding, which slows down the rate of protein production in the cell.
      • Group 8: endonuclease activity
        • Decreased hydrolysis of nucleic acids.
      • Group 9: glucose metabolic process
      • Decreased synthesis of proteins necessary for glycolytic processes (similar to Group 3).
      • Group 10: aromatic compound biosynthetic process
        • Decreased synthesis of proteins involved in the chorismate biosynthesis pathway, which produced energy for the cell.
      • Group 11: intracellular protein transport and intracellular transport
        • Decreased intracellular protein transport.
  • Overall takeaways from investigating the functions of the top 20 genes that exhibited significant expression changes in the Vibrio cholerae microarray experiment:
    • Genes involved in refolding denatured proteins and preventing protein aggregation were down-regulated, whereas the expression of genes that synthesize enzymes that hydrolyze proteins were up-regulated. The pathogenic cell thus exhibit reduced energetic investment in proper protein folding and increased rates of protein degradation. There is additionally a decreased investment in protein transport.
    • Similarly, genes involved in stimulating protein production are being down-regulated in the pathogenic cells. This decreased investment in protein production suggests that these cells are more invested in conserving energy, perhaps in preparation for energy deficits.
    • Genes involved in critical energy-producing metabolic pathways such as glycolysis and chorismate biosynthesis and being down-regulated in the pathogenic cells. At the same time, sugar-dependent active transport is being reduced. Both of these processes signify increased conservation of chemical energy by reducing consumption of carbohydrates in cellular processes. These cell thus appear better equipped to handle energy deficit.

In their 2002 paper "Host-induced epidemic spread of the cholera bacterium", Merrell et al. present the argument that the human gastrointestinal tract provides a suitable growth environment for pathogenic Vibrio cholerae bacteria, which then enter an induced hyperinfectious state. This transition, however, is completely contingent on the survival of V. cholerae in the highly acidic conditions of the human gastrointestinal tract. Thus, ingesting a relatively large quantity of V. cholerae is typically necessary for infection under normal stomach pH conditions (if the pH is lowered, infection is easier). These findings suggest that the acidic environment of the human gastrointestinal tract places a great deal of selective stress on V. cholerae, favoring expression patterns that confer increased resistance to these stresses in pathogenic V. cholerae. The results of my microarray analysis support this hypothesis, as they indicate significantly different expression patterns that confer resistance to the stresses present in the human gastrointestinal tract. Many of these expression patterns revolve around proteins. The pathogenic V. cholerae in the microarray experiment exhibited decreased investment in the correction of protein folding errors and instead favored faster protein degradation. This expression change enables the pathogenic bacteria to better survive in the low pH conditions of the gastrointestinal tract, where proteins would denature rapidly. Instead of constantly trying to refold the denatured proteins, the cell instead slows down protein synthesis on the whole (which also conserves energy) and accelerates degradation of proteins through hydrolysis, which fights against aggregation of denatured proteins. In addition to this, the pathogenic V. cholerae exhibited decreased synthesis of proteins necessary for energy-producing metabolic pathways such as glycolysis and chorismate biosynthesis and decreased sugar-dependent active transport. These changes allow for the conservation of sugars necessary for energy-production, which confers the pathogenic bacteria greater ability to survive in stressed conditions such as in the gastrointestinal tract. In addition, these changes also reduce intake of molecules from the external environment, which could be beneficial to survival in an environment filled with digestive enzymes. Evidently, the pathogenic V. cholerae in this microarray experiment exhibited expression changes that increased their ability to survive in humans' gastrointestinal tracts and thus incite an opportunistic infection.

  • There is one other file you need to save to your journal page. It has a .gmf extension and should be in the same fold as the .gex file that you created with the GenMAPP Expression Dataset Manager. You will need this file to re-open your results in MAPPFinder.

Conclusion

This week, I conducted an analysis of the raw data generated by a real DNA microarray experiment. This process started with statistically analyzing the raw data to yield normalized log fold change values for the different genes and calculate their statistical significance. Subsequently, this data was reformatted and imported into the program GenMAPP, which used the results of this statistical analysis to create a global gene expression profile for the microarray data. By identifying the most significant gene ontology terms in this expression profile, insights about the expression changes in the experimental microarray group were gleaned. These expression changes were applied to the experimental context of the experiment performed by Merrell et al. In doing so, conclusions were drawn about why pathogenic V. cholerae exhibit these expression changes, which I argued was because these pathogenic bacteria require higher resistance to the stresses present in the human gastrointestinal tract to survive.

Links

Assignments Pages

Individual Journal Entries

Shared Journal Entries

Retrieved from "https://xmlpipedb.cs.lmu.edu/biodb/fall2015/index.php?title=Bklein7_Week_8&oldid=3358"