This computer exercise involves searching for matching peptides to MS/MS spectra of shotgun proteomics experiment, as well as the subsequent inference of proteins from these peptides. We will do the search using a freely available search engine called OMSSA, on an example dataset from a yeast lysate. To pass the lab you will have to submit one short report (in .odt, .doc, .docx, .pdf or .rtf format) per group, with answers to the questions, to any of the Teaching Assistants. The deadline is set to one week after the lab has taken place. Please work in pairs, as long as possible, and follow the instructions below:
Here, we will use MS/MS spectra derived from a yeast lysate. Download the MS/MS spectra file here: yeast_spectra.mgf, and put it in a convenient working directory.
A yeast protein database in the FASTA-format can be downloaded from here: database.zip, and is most conveniently placed in the same working directory. Unzip it to be able to use the uniprot_sprot_yeast09_02.fasta-file kept inside, for example by running the following command from your working directory: unzip database.zip
We need two programs to carry out the peptide search. First a Search Graphical User Interface (SearchGUI), to be able to start the search within a graphical setting. And second, the search engine (OMSSA) itself. In addition, we'll use a two additional programs to look at the results (OMSSA-parser) and to infer proteins from the peptides (MassSieve), we'll download these further on in the lab.
Download the SearchGUI from http://compomics.github.io/projects/searchgui.html.
Unzip: Place the archive in your designated working directory and unzip it with the command: tar -xvzf SearchGUI-3.3.9-mac_and_linux.tar.gz
Download OMSSA from ftp://ftp.ncbi.nih.gov/pub/lewisg/omssa/CURRENT/ and extract OMSSA in your working directory.
The task is to run OMSSA using the SearchGUI to identify peptides from the MS/MS spectra. Subsequently, we will identify proteins from these peptides.
1. Get SearchGUI and OMSSA working
Download and extract the SearchGUI and OMSSA as described above:
Under linux, a few permissions may have to be edited in order to make the SearchGUI to work. change the permissions of omssacl in your OMSSA-directory, and makeblastdb in the SearchGUI-3.3.9/resources/makeblastdb/linux/linux_64bit directory to give all users execution rights. For example by running these commands in your working directory:
chmod 777 omssa-2.1.9.linux/omssacl
chmod 777 SearchGUI-3.3.9/resources/makeblastdb/linux/linux_64bit/makeblastdb
From the SearchGUI's directory, open the SearchGUI, for example with the following commands:
cd SearchGUI-3.3.9/
java -jar SearchGUI-3.3.9.jar
If you get an error message mentioning that MS Amanda was not properly installed, you can safely ignore it and also press Cancel on the "Search Engine Settings" dialogue showing up next.
2. Define the required settings of the search
If opening the SearchGUI worked properly, you should now be able to search for matching peptides. But we will first have to specify the parameters of our search.
Under Edit > Software Folders, change the path of OMSSA Location to your unzipped OMSSA directory. Also make sure OMSSA is the only enabled search engine. The other possible search engines, such as X!Tandem, are other search engines which we didn't download here.
Under "Input & Output", select the yeast_spectra.mgf file as the "Spectrum File", and an appropriate "Output Folder".
3. Configure the parameters to adjust the search for our particular experiment
Right below the spectrum file input field, there is a field for search settings. Click add to create a new settings option. Choose any name you prefer and click "Spectrum Matching" to enter the menu for editing the search settings.
For Database, select the uniprot_sprot_yeast09_02.fasta file, if it promts you for Decoys, select "No".
Fixed modifications: 'Carbamidomethylation of C'
Variable modifications: 'Oxidation of M' and 'acetylation of protein n-term'.
Our shotgun proteomics experiment was carried out on yeast proteins digested with trypsin, choose Trypsin as the proteolytic enzyme (protease), with 1 as Max Missed Cleavages. QUESTION 1: why not 0?.
Set the mass tolerance to 0.2 Da (note the unit) in both cases. QUESTION 2: How come we can specify two different mass tolerances?.
We'll leave the ion types and precursor charge bounds as they are. Press ok and save these settings.
4. Run the peptide search
Add a suitable Output Folder and then click on "Start the Search!" button. The search will go in a couple of seconds.
Look in your designated output folder, if everything worked properly, it should contain an archive with a couple of files, including yeast_spectra.omx, omssa_mods.xml, omssa_usermods.xml. The OMX-file contains the peptides matches, and the other two files the OMSSA modifications and the user modifications file. Extract the archive to a convenient location and rename the two modification files to mods.xml and usermods.xml, as this is necessary for the next step. We will look closer at these files, and try to infer what proteins they represent in the next step.
5. Watch the peptide search results
The peptide search is now completely finished and three files have been created in your output directory: First, an OMSSA XML-file (.omx) listing all the peptide matches, and second, two files (mods.xml and usermod.xml) with OMSSA's and our own modification settings, respectively.
To look at the results we use the OMSSA-parser. Download omssa-parser-1.9.0.zip from http://compomics.github.io/projects/omssa-parser.html and unzip it in your working directory. (Command: unzip omssa-parser-1.9.0.zip).
Open the parser from its' extracted directory, using the command java -jar omssa-parser-1.9.0.jar
You will be asked for your peptide search results file and the two modification files, fill in these fields.
Now, you'll see a list of all fragmentation spectra, some of which have been matched to one or many peptides in the database fasta-file. Click on a spectrum to see it, and its' matched peptides. Click around a little bit to get a feeling of it.
QUESTION 3: For each spectrum, how are the peptide matches ranked originally? (Assuming you have not resorted them...)
QUESTION 4: What do the unidentified spectra seem to have in common?
6. Protein identification
After searching for peptides, download the "Latest Development Binary" of MassSieve from http://www.ncbi.nlm.nih.gov/staff/slottad/MassSieve/. Unzip to a directory called MassSieveDev (Command: unzip -d MassSieveDev masssieve.zip), in your working directory.
Open MassSieve from the newly created MassSieveDev directory with the following command:
java -jar MassSieve.jar
A new window will open, with only a few menus available at the top, to process the peptide identifications results from OMSSA, start a new "experiment" with "File > New Experiment".
The next step is to add our data, to do this, choose "File > Add Search Results". This step infers proteins thought to be in our yeast sample from the peptide identifications.
The right side of the screen shows the proteins matches, clicking on a protein row can show a "Cluster Graph" of the manner in which the protein is identified. Some proteins are identified by many peptides, some by just a single peptide match.
Scrolling down to see the rest of the protein identifications will reveal some of the problems with protein inference, indicated by the word in the "Parsimony Type" column. Look at the "Cluster Graph" of a few proteins, with different "Parsimony Types".
QUESTION 5: What seems to be the difference between the proteins identified with different parsimony types?
QUESTION 6: Can you find what important information about the protein matches that is missing in this program?
1. 08:25 Feb 9, 2011:
There might be some problems with limited space on your login accounts (the lab requires ~200 MB). A first attempt to solve this could be by working in the /tmp folder, which is the temporary folder on your computer (cd /tmp).
A second attempt to solve this could be by working on a plugged-in USB-stick.
2. 08:58 Feb 9, 2011:
To avoid problems when downloading extracted-files to the USB-stick, choose OPEN instead of SAVE after clicking on the link. Then extract the files directly to your USB-stick.
3. 09:45 Feb 9, 2011:
An error message during the OMSSA peptide search might be due to a truncated yeast_spectra.mgf-file, compare the end of the file on your computer with the file you copied from this website. Replace it if you see any differences.