PNNL

Abstract

Peptide identification following tandem mass spectrometry is usually achieved by searching for the best match between the mass spectrum of an unidentified peptide and model spectra generated from peptides in a sequence database or by piecing together a feasible path in a graph developed from experimental data. The first of these identification methodologies will be successful only if the peptide under investigation belongs to an available database and the matching algorithm is mathematically sound; the second if the experimental spectral information is very close to perfect. Our objective is to develop a new framework for this problem that leads to alternative solution methodologies better suited to deal with features commonly found in actual MS/MS spectra.