THE SCIENTIST STAFF
Mass spectrometers can quickly spit out spectra of an organism’s complete complement of proteins, known as the proteome. Shotgun experiments, also known as data-dependent analyses, aim to ID as many sample proteins as possible. “This is a great method to establish a protein inventory of a cell or tissue,” says Ruedi Aebersold of the Swiss Federal Institute of Technology in Zurich. “But it is less well suited to compare precisely the abundance and presence of specific proteins across many samples.”
That’s where a targeted mass spectrometry technique known as selected reaction monitoring (SRM) can come in handy. It’s a tried-and-true method that has been around for decades. “With the SRM approach, you program the machine in advance to look for a specific peptide that you know is unique to the protein you’re interested in,” says Conrad Bessant, a professor of bioinformatics at Queen Mary University of London in the U.K. If that peptide is present, the mass spectrometer smashes it into pieces and quantifies a specific fragment of that peptide. “It will do this with a high level of reproducibility and a high degree of quantitative accuracy,” Aebersold says.
While advanced mass spec–based methods for targeted protein analysis have been developed in recent years (see “Moving Target,” The Scientist, June 2014), approaches based on SRM are the most mature and widely used. Of course, an SRM approach covers a much smaller subset of the proteome than a shotgun strategy, and it requires more up-front legwork to design an SRM assay. But by carefully programming the mass spectrometer to perform sequential rounds of SRM, a strategy known as multiple-reaction monitoring (MRM), researchers can accurately detect and quantify hundreds of proteins in a sample at a time. And the potential pitfalls of SRM—that it can be slow or cumbersome—can be addressed with bioinformatics and careful planning.
To develop an SRM or MRM assay, researchers first need to identify peptides that map only to their protein of interest and are not found in any other protein that could be present in their sample. They also need to choose at least one fragment ion to serve as a marker for each particular peptide ion produced in the mass spec. Then the instrument can be programmed to look specifically for the mass-to-charge (m/z) ratios of the peptides and fragment ions of interest. Each particular combination of peptide and fragment ion is known as a transition, and must meet a number of criteria. For example, some peptides have a tendency to ionize poorly, and would thus be unsuitable for mass spectrometry. Others may be subject to posttranslational modifications, which can interfere with detection and quantification.
It is possible to manually design SRM and MRM assays, but the process can be challenging and time-consuming. Here, The Scientist profiles five free computational tools that can streamline the process.
What It Does: MRMPath is an online tool that predicts suitable peptides for MRM assays based on protein sequence. Users first enter the accession number or sequence of their protein of interest. Alternatively, users can choose a particular biological pathway and species, and MRMPath generates a list of all the proteins involved in the chosen pathway using information stored in the Kyoto Encyclopedia of Genes and Genomes database (KEGG; www.genome.jp/kegg/). Users then select one of five different enzymes commonly used in mass spectrometry-based proteomics to produce peptide fragments, and the program computationally digests the selected protein into smaller pieces. MRMPath generates a list of 7- to 25-amino-acid-long peptides, automatically excluding any that contain methionine or cysteine amino acid residues, which are susceptible to oxidation and can interfere with analysis by mass spectrometry.
The program also provides the predicted m/z values of each peptide and its fragment ions. From there, users can click the “BLAST” button, and MRMPath will search protein sequence databases compiled by the National Center for Biotechnology Information (NCBI; blast.ncbi.nlm.nih.gov) to identify peptides that are unique to the protein of interest. MRMPath helps users to optimize their SRM and MRM assays for specificity, says Stephen Barnes of the University of Alabama at Birmingham whose lab developed the tool. “What we’re really trying to avoid is having a mass-spec experiment that is not unique to a protein or to a species.”
- Users can choose from among hundreds of cellular pathways and disease processes, and from among thousands of different species.
- The MRMPath website hosts two additional tools: MRMMut enables users to quickly determine whether peptides have known mutations, which can interfere with an MRM assay; MRMSpace helps users to determine whether any other proteins in a limited number of species can give rise to peptides and fragment ions with m/z values that match those of a transition selected for a particular MRM assay.
- Because the results are based on theoretical predictions, there’s no guarantee that the suggested peptides and fragment ions will be detectable by a mass spectrometer.
LAPTOP: ©GRAPHICSDUNIA4YOU/ISTOCKPHOTO.COMWhat It Does: MRMaid is a Web-based application developed by Bessant’s lab that suggests SRM transitions based on the peptides and fragment ions that have been detected previously in shotgun proteomics experiments stored in the public data repository known as PRIDE (PRoteomics IDEntifications; www.ebi.ac.uk/pride). MRMaid users simply choose their species of interest (currently limited to human and Arabidopsis thaliana) from a drop-down menu and enter the name or accession number of one or more proteins into the search box. Then the program uses data from PRIDE to generate a list of potential transitions to monitor in an SRM or MRM experiment.
MRMaid also provides information that can aid in the design of a targeted assay, such as the peptide’s retention time, the m/z values for the peptide and fragment ion, and the average intensity of the ion. Each transition in the list is assigned a so-called peptide score, which indicates its suitability for SRM based on the quality of the peptide’s available fragmentation spectra.
- Easy-to-use visual interface, says Max Wong, a user at the University of Cambridge in the U.K.
- Users can search for several different proteins at once, which can be useful when designing complex multiprotein assays.
- Interactive; the list can be sorted or filtered according to several different criteria.
- Doesn’t provide information on potential interferences—other transitions in the sample that have similar m/z values to the transition of interest
- Doesn’t filter out peptides subject to posttranslational modifications or those likely to be miscleaved by proteases
What It Does: PeptidePicker is a Web-based application that integrates information from a variety of different data sources to compile a list of optimal peptides for an SRM/MRM assay. Users enter a protein accession number, choose their species of interest (human or mouse only), and specify a minimum and maximum length of peptides (the default setting length is 7–20 amino acids). If desired, users can also choose to limit the program’s focus only to peptides that satisfy certain criteria, such as only those peptides that exist in all known isoforms of a protein, or peptides that lack known or potential posttranslational modifications.
PeptidePicker first computationally digests the protein into peptides with the enzyme trypsin, and excludes any peptides that are not unique to the proteome of the specified organism. Then the program generates a list of potential peptides ranked according to a so-called v-score, which indicates their suitability for SRM/MRM assays according to the user’s optimized criteria.
In generating the v-score, PeptidePicker automatically queries large proteomics repositories, including PRIDE, the Global Proteome Machine database (GPMDB; gpmdb.thegpm.org), and PeptideAtlas (www.peptideatlas.org), and rewards peptides that have been observed previously by mass spectrometry. The program also searches protein-specific databases, including UniProtKB (www.uniprot.org/uniprot/) and the NCBI’s database of single nucleotide polymorphisms (dbSNP; www.ncbi.nlm.nih.gov/SNP/), and penalizes peptides that have any known sequence variations. “This tool combines our knowledge—years and years of experience of how the selection should be done—with what is known and described in databases,” says PeptidePicker developer Christoph Borchers, director of the Proteomics Centre at the University of Victoria in British Columbia, Canada.
- Results are displayed in an easy-to-comprehend format.
- Users can control several of the peptide selection criteria.
- Handles only one protein at a time
- Only suggests suitable peptides; users will need to use other tools or perform experiments to identify fragment ions
THE SCIENTIST STAFFWhat It Does: SRMAtlas is a database of SRM transitions and mass spectra derived from synthetic peptides created for SRM assay development and analyzed on the types of mass spectrometers most frequently used for SRM/MRM experiments—triple quadrupole, Orbitrap, and quadrupole–time-of-flight instruments. Users can search SRMAtlas using a protein accession number or peptide sequence, or by uploading a text file containing a list of protein names or peptide sequences. “We made at least eight peptides for every protein, and then measured them all using a number of mass spectrometers from different manufacturers,” says Rob Moritz of the Institute for Systems Biology in Seattle, whose lab created SRMAtlas together with Aebersold’s team in Zurich. “It’s sort of like a one-stop resource in which essentially all proteins from an organism have at least one assay or more per protein.”
- The suggested transitions are derived from synthetic peptides and should be more reliable than those based on shotgun proteomics data, says Andrei Drabovich, a user at the University of Toronto in Canada.
- The SRMAtlas also contains a database called PASSEL, which catalogs “validated” SRM assays that have been used successfully in the context of a particular sample.
- The spectra stored in the SRMAtlas are also useful for the analysis of data generated using the newer targeted proteomics techniques.
- Limited coverage—currently only contains data for the complete proteomes of Saccharomyces cerevisiae and Mycobacterium tuberculosis. However, Moritz says his team has finished collecting data for the complete proteomes from human, mouse, rat, and other species, and hopes to release the results later this year (a trial version containing a portion of the human atlas is already available on the site).
- Doesn’t predict potential interferences
What It Does: Skyline is a Windows-based desktop application developed by Michael MacCoss’s lab at the University of Washington in Seattle for designing SRM/MRM assays and analyzing the resulting mass spectra. Skyline helps users select peptides and transitions based on mass spectra from their own experiments or imported from public proteomics data repositories. Users first import the sequence of one or more proteins of interest into Skyline, and the program computationally digests the proteins into peptides. Skyline can automatically exclude peptides that don’t adhere to certain user-defined criteria, such as sequence and length. If desired, Skyline can also suggest or refine a list of potential peptides using imported spectra from previous studies. Users then analyze the suggested peptides by mass spectrometry, and the program can help determine which transitions to monitor based on signal intensity and specificity.
- Very nice user interface, and a wide range of useful features and functionalities, says Eduard Sabidó, head of the proteomics unit at the Center for Genomic Regulation in Barcelona, Spain
- To aid the selection of unique peptides, users can create and import into Skyline a so-called “background proteome” file containing sequence information for all of the proteins expected to be in a given sample, says Christine Carapito, a user at the University of Strasbourg in France.
- Also useful for analyzing data obtained using newer targeted proteomics techniques
- Several tutorials and training videos available on the Web
- Runs on Windows only. But it can work just fine on Apple and Linux machines running Microsoft Windows in a virtual environment, says Chris Colangelo, director of the Protein Profiling Resource at Yale University.