Feature Selection for Tandem Mass Spectrum Quality Assessment

In the literature, hundreds of features have been proposed to assess the quality of tandem mass spectra. However, some features may be nearly irrelevant, and thus the inclusion of these nearly irrelevant features may degenerate the performance of quality assessment. This paper introduces a two-stage support vector machine recursive feature elimination (SVM-RFE) method to select the most relevant features from those found in the literature. To verify the relevance of the selected features, the classifiers with the selected features are trained and their performances are evaluated. The out performances of classifiers with the selected features illustrate that the set of selected features is more relevant to the quality of spectra than any set of features used in the literature.