PISV: A novel algorithm for peptide identification using spectrum vector

Peptide analysis by tandem mass spectrometry (MS/MS) is an important method for protein identification. The commonly used and most mature method of peptide identification is the database search algorithm, which searches the experimental spectrum against peptide database to find out the peptide that gives the best interpretation of the spectrum. The widely-varying intensity range and the presence of many unknown peaks are common challenges for interpretation of mass spectra, and the alignments between observed peaks within mass spectrum and theoretical fragments of peptide are critical tasks for peptide identification algorithms. HMM_Score is a successful database search algorithm for MS/MS data, but it fails to effectively solve these problems mentioned above. To address these issues, we developed an accurate database search algorithm for peptide identification based on cross-assignment between observed peaks and theoretical fragments. In addition, a more efficient strategy used in our algorithm greatly reduces the time cost for database searching. We evaluated our algorithm on two data sets publicly available, and the results show that our algorithm outperforms HMM_Score.