Hotel Estoril Eden, Monte Estoril,
Portugal
5-8 October 2005

 

 
 

Text Box: Participants
Text Box: Programme


A
comparative study of computational tools for the analysis of LC-MS data

Olga Vitek1, A. Garbutt1, N. Zhang1, X.-J. Li1, E. Yi1 and R. Aebersold1,2
1
Institute for Systems Biology, Seattle, WA, USA
2Institute for Molecular Systems Biology, Zurich, Switzerland

Liquid chromatography-mass spectrometry (LC-MS) offers an exciting opportunity for quantitative proteomics and for discovery of new molecular biomarkers. LC-MS data are in the form of 2-dimensional spectra. Peaks in the spectra represent charged peptide ions, and their volumes are related to the abundance of the peptides in the original sample. Thus the peaks can potentially be used to simultaneously quantify the abundance of many proteins in a cell across multiple conditions.

Although the LC-MS data appear somewhat similar to the data obtained via a
microarray experiment, their statistical analysis presents additional challenges. In particular, the number of peptides, their identities and their locations in the spectra are unknown. Before we can proceed with a study of differential expression, one needs to: 1) distinguish peptide peaks from noise; 2) match peptide peaks across spectra from multiple runs; 3) determine the volumes of the peaks. Three computational tools, namely SpecArray, msInspect and massAnalyzer have recently been developed for these steps. Here we compare the performance of the tools on the basis of an experimental test data set.