A Machine Learning Approach to Mass Spectra Classification with Unsupervised Feature Selection

Authors:
Michele Ceccarelli;Antonio D'Acierno;Angelo Facchiano
Affiliations:
Department of Biological and Environmental Sciences, University of Sannio, Benevento, Italy 82100 and Research Center on Software Technologies, University of Sannio, Benevento, Italy 82100 and Bio ...;Institute of Food Sciences, Italian National Research Council, Avellino, Italy;Institute of Food Sciences, Italian National Research Council, Avellino, Italy
Venue:
Computational Intelligence Methods for Bioinformatics and Biostatistics
Year:
2009

Citing 14
Cited 0

A training algorithm for optimal margin classifiers

COLT '92 Proceedings of the fifth annual workshop on Computational learning theory
The nature of statistical learning theory

The nature of statistical learning theory
Scale-Space Theory in Computer Vision

Scale-Space Theory in Computer Vision
A Tutorial on Support Vector Machines for Pattern Recognition

Data Mining and Knowledge Discovery
Kernel Methods for Pattern Analysis

Kernel Methods for Pattern Analysis
Particle swarm optimization for analysis of mass spectral serum profiles

GECCO '05 Proceedings of the 7th annual conference on Genetic and evolutionary computation
Reproducibility of SELDI-TOF protein patterns in serum: comparing datasets from different experiments

Bioinformatics
Sample classification from protein mass spectrometry, by 'peak probability contrasts'

Bioinformatics
An introduction to ROC analysis

Pattern Recognition Letters - Special issue: ROC analysis in pattern recognition
Peak selection from MALDI-TOF mass spectra using ant colony optimization

Bioinformatics
Improved model-based, platform-independent feature extraction for mass spectrometry

Bioinformatics
Independent component analysis for the extraction of reliable protein signal profiles from MALDI-TOF mass spectra

Bioinformatics
Comparing support vector machines with Gaussian kernels to radialbasis function classifiers

IEEE Transactions on Signal Processing
A comparison of methods for multiclass support vector machines

IEEE Transactions on Neural Networks

Quantified Score

Hi-index	0.00

Visualization

Abstract

Mass spectrometry spectra are recognized as a screening tool for detecting discriminatory protein patterns. Mass spectra, however, are high dimensional data and a large number of local maxima (a.k.a. peaks ) have to be analyzed; to tackle this problem we have developed a three-step strategy. After data pre-processing we perform an unsupervised feature selection phase aimed at detecting salient parts of the spectra which could be useful for the subsequent classification phase. The main contribution of the paper is the development of this feature selection and extraction procedure grounded on the theory of multi-scale spaces. Then we use support vector machines for classification. Results obtained by the analysis of a data set of tumor/healthy samples allowed us to correctly classify more than 95% of samples. ROC analysis has been also performed.