Find Statistically Significant Modifications with Limelight

This tutorial covers how to use Limelight to compare the results of two searches (treated and untreated) to find the modifications with the most statistically significant differences (in terms of their spectral counts) between the two conditions. This is the method employed in the following publication:

Discovery and Visualization of Uncharacterized Drug-Protein Adducts Using Mass Spectrometry. Riffle M, Hoopmann MR, Jaschob D, Zhong G, Moritz RL, MacCoss MJ, Davis TN, Isoherranen N, Zelter A. Anal Chem. 2022 Mar 1;94(8):3501-3509. doi: 10.1021/acs.analchem.1c04101

This tutorial will assume the user has completed our Magnum Pipeline Tutorial to generate a treated.limelight.xml file, and has completed the optional step to produce the untreated.limelight.xml file.

Upload Data to Limelight

Although uploading data is covered in our Magnum Pipeline Tutorial, we will repeat those steps here. If you have already uploaded your treated.limelight.xml and untreated.limelight.xml files to Limelight, skip this step.

Note

If you are using Windows, this step is greatly simplified by first copying your Limelight XML files to a Windows filesystem drive, such as C:\. To copy treated.limelight.xml and untreated.limelight.xml to C:\data_directory\ you would enter the following into your terminal.

cp ~/my-project/treated.limelight.xml /mnt/c/data_directory/treated.limelight.xml
cp ~/my-project/untreated.limelight.xml /mnt/c/data_directory/untreated.limelight.xml

# Optional: You must upload your mzML files to view spectra. Copy to C:\ for upload later
cp ~/my-project/treated.mzML /mnt/c/data_directory/treated.mzML
cp ~/my-project/untreated.mzML /mnt/c/data_directory/untreated.mzML

/mnt/c/ corresponds to your C:\ drive. /mnt/d/ corresponds to your D:\ drive, and so on.

Steps to upload your data to Limelight:

1. Log into Limelight

2. Navigate to an existing project or create a new project.

3. Scroll down to the Upload Data section and click the arrow on the left to expand the section.

   

4. Click the Import Limelight XML File button. You should see the following dialog appear:

   

5. Enter a description of this run. E.g., Treated.

6. Click on the +Add Limelight XML File link and select your treated.limelight.xml on your computer.

7. (Optional) Click on the +Add Scan File link and select your treated.mzML file on your computer. This is required if you would like to view spectra associated with peptide identifications.

8. Click the Submit Upload button to submit your data to Limelight.

9. Repeat steps 4-6 for your unreated.limelight.xml and, optionally, your untreated.mzML file.

After several minutes, refresh the page and your searches should appear under the Explore Data section of the project page.

Navigate to the Mod View Page

Once your data are imported, the searches will appear under the Explore Data section:

You will see your two searches listed here. To find the modifications with significantly different spectral counts between the searches, click the check box next to each search and click the Compare Mod View button:

Change Visualization Options

By default the mod page will show the following:

The heat map shows the spectral count for all PSMs with a given mod mass in each of the two searches. In this example, the spectral counts for 0, 16, and 57 are drowning out the signal for the open modification masses, which will have a modification mass of 60 or higher.

To restrict the mod masses to those with a mass of 60 or higher, change the minimum mod mass to 60 and click the Update Visualization button, as depicted below:

The page should update to the following:

The heat map is more informative. Note the bands that appear near 470 in the treated sample, but not the untreated sample.

Run the Report

To run a statistical analysis comparing the spectral counts for mod masses in the two searches, click the View ZScore Report link below the data visualization:

This will compare the ratio of PSMs that have a given mod mass to all PSMs in each search using a test for proportions and produce the following report:

This report is ordered by the magnitude of the Z-score. Note that the modification masses 469, 470, and 471 have the most significant Z-scores. A negative Z-score in this case denotes enrichment in the treated sample.