Where academic tradition
meets the exciting future

You are here: TUCS > PUBLICATIONS > Publication Search > Filtering Bad Quality Tandem M...

Filtering Bad Quality Tandem Mass Spectra Prior to Protein Identification

Jussi Salmi, Robert Moulder, Jan-Jonas Filén, Olli S. Nevalainen, Riitta Lahesmaa, Tuula A. Nyman, Tero Aittokallio, Filtering Bad Quality Tandem Mass Spectra Prior to Protein Identification. TUCS Technical Reports 729, Turku Centre for Computer Science, 2005.

Abstract:

An important task in proteomics is the identification of peptides from
tandem mass spectrometry data. Peptide identification is usually
accomplished by using database search programs which match the experimental
spectral sequences with theoretical sequences. As the search results
frequently contain false matches, they need to be validated manually. Many
spectra do not contain peptide sequences and could be discarded before the
time-consuming search process. We present a filtering method which
addresses these problems by classifying the spectra into two classes: (i)
the spectra that are unlikely to produce valid matches and (ii) the
presumably valid spectra. The filter is based on 9 spectral features, which
measure different characteristics of a spectrum. The discriminability of
these features is investigated in conjunction with a machine learning
algorithm using a training set of instances and tested on real-life data
sets. The results show that when removing about half of the spectra, the
number of protein identifications dropped 0-25% depending on the material,
but the amount of time spent on the identification process was sharply
reduced.

Only abstract available. Not in public distribution.

BibTeX entry:

@TECHREPORT{tSaMoFiNeLaNyAi05a,
  title = {Filtering Bad Quality Tandem Mass Spectra Prior to Protein Identification},
  author = {Salmi, Jussi and Moulder, Robert and Filén, Jan-Jonas and Nevalainen, Olli S. and Lahesmaa, Riitta and Nyman, Tuula A. and Aittokallio, Tero},
  number = {729},
  series = {TUCS Technical Reports},
  publisher = {Turku Centre for Computer Science},
  year = {2005},
  keywords = {MS/MS spectra analysis, classification, protein identification},
  ISBN = {952-12-1651-4},
}

Belongs to TUCS Research Unit(s): Algorithmics and Computational Intelligence Group (ACI), Biomathematics Research Unit (BIOMATH)

Edit publication