Related Experiment Videos
Peptide mass fingerprinting peak intensity prediction: extracting knowledge from spectra
Steven Gay1, Pierre-Alain Binz, Denis F Hochstrasser
1Swiss Institute of Bioinformatics, Geneva, Switzerland.
Proteomics
|November 8, 2002
Summary
This study introduces datamining methods to predict peptide peak intensities in mass spectrometry. These novel approaches improve the accuracy of peptide mass fingerprinting identification tools in proteomics.
Area of Science:
- Proteomics
- Analytical Chemistry
- Bioinformatics
Background:
- Matrix-assisted laser desorption/ionization-time of flight mass spectrometry (MALDI-TOF MS) is crucial in proteomics.
- Current peptide mass fingerprinting (PMF) tools lack accuracy and efficiency due to limited use of peak intensity data.
Purpose of the Study:
- To develop accurate and efficient datamining methods for peptide mass fingerprinting (PMF) identification.
- To correlate peptide physicochemical properties with their peak intensities in MALDI-TOF MS spectra.
Main Methods:
- Applied standard datamining classification (C4.5) and regression (M5') methods.
- Utilized a dataset from 157 proteins analyzed via PMF experiments.
- Analyzed correlations between experimental and theoretical peptide masses and intensities.
Main Results:
- C4.5 method achieved 88% accuracy in classifying theoretical peaks.
- M5' regression model showed a 0.6743 correlation coefficient for normalized peak intensities.
- Developed decision and model trees for direct PMF prediction and identification.
Conclusions:
- Datamining methods can effectively predict peptide peak intensities based on physicochemical properties.
- This approach lays the foundation for improved accuracy in PMF analysis.
- The methodology is extendable to other mass spectrometry techniques and datasets.