Galaxy:  Portal | History | About Galaxy | Example queries | Help | FAQ | Contact us

Example Queries

TABLE OF CONTENTS

Note:  these instructions assume you are a new user with no queries in your history. If you already have some queries, then the new ones will be numbered differently than shown here.

Example from our poster at the 2005 CSHL Biology of Genomes meeting

This example uses the KA/KS ratio to help determine which predicted coding regions are likely to be real. Then an example exon is selected and a sliding profile graph is shown for that exon.

Steps:

  1. Start by getting the experimentally verified coding regions from the UCSC Table Browser.
    1. From the Galaxy portal, follow the link to the table browser.
    2. Make sure clade is "Vertebrate", genome is "Human", and assembly is "May 2004". If it is not selected, select "Gene and Gene Prediction Tracks" under group. Next select CCDS for the track. In the position box type "chr22".
    3. Click the "create" button next to the filter. On the page that comes up type "+" in the box next to strand, and then click the "submit" button.
    4. The Galaxy output format should be preselected for you, so now click the "get output" button.
    5. On the next page select "Coding Exons" and then click the "Send query to Galaxy" button. This will add the experimentally verified coding regions to your history page.
  2. Next get the predicted coding regions from the UCSC Table Browser.
    1. Follow the link at the top of the page to the portal page.
    2. From the portal follow the link to the table browser.
    3. The table browser will remember your selections from the previous query so only the track will need changed to "SGP Genes". Changing the track will clear the filter.
    4. Create the filter again the same as before.
    5. Click the "get output" button.
    6. Select "Coding Exons" and then click the "Send query to Galaxy" button. This will add the predicted coding regions to your history page.
  3. Now get the predicted regions that have not been experimentally verified.
    1. Select your previous queries "1" and "2". If it is not selected for you, select "Perform operations like intersection, etc.". Then click the "Go" button.
    2. Select "Subtraction", then make the selections so that it reads "query #2 minus query #1, removing whole regions". Then click the "Go" button. The predicted but not verified regions have now been added to your history.
  4. Standard KA/KS works best on longer exons, so remove the very short ones.
    1. Select query "3" and "Perform operations like intersection, etc.". Click the "Go" button.
    2. Because only one query was selected, different operations are available this time. Select "Restrict region size" and type "150" in the first text box. This will give us only regions that are 150 bp long or more. Click the "Go" button.
  5. KA/KS runs on alignments, but the query results are regions in BED format. So next, fetch the alignments that correspond to these regions.
    1. Select query "4" and "Run analysis tools". Click the "Go" button.
    2. Select "Alignment retriever" and next to it "hg17 - mm5". This will give the Human-Mouse pairwise alignments that overlap the coding regions from query 4. Click the "Go" button.
  6. Now compute the standard KA/KS ratio on the alignments.
    1. Select query "5" and "Run analysis tools". Click the "Go" button.
    2. Select "KA/KS Estimation", and click the "Go" button.
    3. Run the Standard KA/KS analysis, leaving the defaults for genetic code and weighting. Select "Return Output to Galaxy" and click the "Get Standard Ka/Ks" button.
    4. You will see a message about verifying the selections, then you will be forwarded to Galaxy, with the standard KA/KS ratios added to the history page.
  7. Find the regions that are likely to be real coding regions, as in PubMed 12552088.
    1. Select query "6" and "Run analysis tools". Click the "Go" button.
    2. Select "Limit query based on score value", and click the "Go" button.
    3. Select "Ka/Ks" for the score name. Type "0.5" in the text box next to "score less than or equal to". Click the "Go" button. This adds the regions with a low KA/KS ratio to the history page.

The remaining steps select a particular region and generate a sliding profile graph. While not necessary to answer the initial question, they still demonstrate the power and usefulness of Galaxy.

  1. Our first attempt at limiting the KS was done on query "6" using score name "Ks" and value "1". This did not prevent peaks from showing up in the final graph, which was the goal of these additional steps. It was left in the history to show the difference in the number of regions affected (compare to query 12).
    1. Select query "6" and "Run analysis tools". Click the "Go" button.
    2. Select "Limit query based on score value", and click the "Go" button.
    3. Select "Ks" for the score name. Type "1" in the text box next to "score less than or equal to". Click the "Go" button. This adds the regions with a low KS value to the history page.
  2. Get the alignments for the likely-real coding regions.
    1. Select query "7" and "Run analysis tools". Click the "Go" button.
    2. Select "Alignment retriever" and next to it "hg17 - mm5". This will give the Human-Mouse pairwise alignments that overlap the coding regions from query 7. Click the "Go" button.
  3. Get the KS scores for the sliding window profile.
    1. Select query "9" and "Run analysis tools". Click the "Go" button.
    2. Select "KA/KS Estimation", and click the "Go" button.
    3. Run the Sliding Window KA/KS analysis with default settings, except choose "Return Output to Galaxy" instead of "Display Output". Click the "Get Sliding Window Ka/Ks" button. This also returns the KS values for the sliding window.
  4. Find all regions with KS greater than or equal to 1.
    1. Select query "10" and "Run analysis tools". Click the "Go" button.
    2. Select "Limit query based on score value", and click the "Go" button.
    3. Select "Ks" for the score name. Type "1" in the text box next to "score greater than or equal to". Click the "Go" button. The regions with KS of 1 or higher are added to your history.
  5. Find the regions where KS does not go over 1.
    1. Select your previous queries "7" and "11". If it is not selected for you, select "Perform operations like intersection, etc.". Then click the "Go" button.
    2. Select "Subtraction", then make the selections so that it reads "query #7 minus query #11, removing whole regions". Then click the "Go" button. The regions where KS stays below 1 are added to your history.
  6. Select one region from the results for the graph.
    1. Select query "12" and "Get output". Click the "Go" button.
    2. Select the "Raw result file" option and click the "Go" button to view the results. Copy the coordinates of one of the regions to the clipboard using your mouse (our poster used the first one). Use the browser's Back button to return to the history page, and click "Refresh" to re-sync it. Then select query "12" and "Perform operations like intersection, etc.". Click the "Go" button.
    3. Select "Restrict position" and paste in the coordinates, editing them to the demonstrated format. Click the "Go" button. This will get this single result by itself.
  7. Get the alignment for this region.
    1. Select query "13" and "Run analysis tools". Click the "Go" button.
    2. Select "Alignment retriever" and next to it "hg17 - mm5". Click the "Go" button.
  8. Generate the graph.
    1. Select query "14" and "Run analysis tools". Click the "Go" button.
    2. Select "KA/KS Estimation", and click the "Go" button.
    3. Run the Sliding Window KA/KS analysis with default settings. Click the "Get Sliding Window Ka/Ks" button.
    4. Select query "15" and "Get output". Click the "Go" button.
    5. Select "Raw result file" and click the "Go" button.

The graph should look similar to figure 4c. in the poster [PDF, 629 Kb].