Wikiomics:Protein mass spectrometry

From OpenWetWare
Revision as of 11:41, 11 December 2007 by Darek Kedra (talk | contribs) (+test data sets, tutorial)
Jump to navigationJump to search

Protein mass spectrometry can be divided into:

  • identification of proteins/peptides
  • quantification

A good introductory tutorial from USC Computational Biology group is here.

Protein/peptide identification

Peptide Mass Fingerprinting (PMF) or (MS)

Old method, superseded by MS/MS

  • algorithms:
  • caveats
    • no sequence information
    • journals started to require that at least one peptide of a protein identified by PMF should be confirmed by MS/MS

Peptide fragment fingerprinting (PFF) or (MS/MS)

  • algorithms (other/new/experimental):
    • Spectrum Mill $$$
    • MASPIC
      • this paper claims 5-15% more confident hits than Sequest: [1]
    • InsPecT A new variable mods search from Pevzner & Tanner @UCSD (free?)
  • filtering of the results
    • Trans Proteomic Pipeline [2] (free?)
    • DTASelect it seems to be in a semi-frozen state (free for nonprofit but requires signed MTA)


Protein databases

Use (if possible):

  • IPI International Protein Index
  • always use target-decoy database (i.e. concatenated: human_IPI + reversed_human_IPI)
  • decoy databases creation methods:
    • protein reversal (simple to perform. does not scramble fortunately quite rare palindromic sequences)
    • peptide pseudo-reversal (used in Sorcerer by Sage-N Research)
      • MEGGAYGAGKAGGAFDPYTL => (trypsin digest, used Ms-Digest) MEGGAYGAGK AGGAFDPYTL => GAGYAGGEMK-ALTYPDFAGG (each peptide reversed, but trypsin digestion site preserved -> guessed from the Elias 2007 paper)
    • shuffled
      • MEGGAYGAGKAGGAFDPYTL => FYAGADEAGMGTYKGGAGLP (used SMS, results differ each time) -> recommended by EBI ppl
    • random (i.e. creating database of random proteins based on frequency of AA in source fasta file)
  • to create decoy database use DBToolkit free java standalone

Modification databases

  • Unimod (> 500 natural + labels)
  • Delta Mass A Database of Protein Post Translational Modifications (in vivo)
  • RESID detailed descriptions of > 400 modifications

Peptide Tag Searching

"Designed to characterize peptides with mutations or unexpected post-translational modifications." (from Popitam page)

  • GutenTag free for non-profit, MTA required. Assigns fewer peptides than Sequest but with fewer false positives. Occupies a middle ground between mainstream search algorithms and de novo sequencing.

de novo sequence determination algorithms

Spectral matching

The idea is that if one can match spectrum of an unknown peptide to a very similar MS/MS spectrum in a database with a determined sequence/annotation then one can annotate unknown peptide in a process similar to orthologue annotation in protein sequence databases. Caveat: bad annotations will also get propagated.

  • P3 (server) from Global Proteomics Machine (free)
  • SpectraST from ISB, Seattle (not as many species/options as P3). Ca 500x faster than Sequest on the same set.

lecture notes by Henry Lam from ISB

  • BiblioSpec from MacCoss lab. (free for non-profit, online licence)
    • command line only

Spectral libraries available here@PeptideAtlas

Protein quantification

  • approaches
    • isotopic labeling (ICAT, ITRAQ, SILAC, 18O- or 15N-labeling)
    • label-free methods
  • software

"calculates the relative abundances of proteins and the corresponding confidence intervals from ICAT-type ESI-LC/MS data"

    • MSQuant Parser for Mascot results for quantitation (Windows only)


  • Trans Proteomic Pipeline (TPP) most popular, included in Sorcerer from Sage-N Research. Windows/Cygwin/Perl or Linux based.
  • Open-MS German, C++ based
  • Sorcerer $$$ FPGA-based fast hardware solution for SEQUEST & Tandem searches with TPP on top of it.

File formats

  • .bdx from Bruker Daltonics
  • .dta SEQUEST/Thermoelectron. Two versions:
    • single pectra
    • multiple spectra concateneted in one file
  • .mgf multiple spectra, Mascot (Matrix science)
  • mzXML (used by Trans Proteomics Pipeline)
  • mzData (standard set by HUPO Proteomics Standard Initiative)

Spectrum datasets

Good for testing programs:

Web sites


For a good review of programs and aspects of protein identification by mass spectrometry see:


Other tools to be sorted out

Ortholog searches using sequence tags/ambigous sequence

$$$ programs


Needs to be sorted out.


  • DeNovoID web
  • SPIDER (PDF) de novo + homology search in other species based on a set of tags
  • OpenSea (HTML) Java program available from authors
ModifiComb (HTML) (available from authors?)
  • MODi web server for PTMs discovery
  • SILVER view your spectra with LOD scores