Blind identification of post-translational modifications via dynamic time warping model

Post-translation modification (PTM) plays a significant role in the understanding of protein functions. However, PTM identification is a very challenging problem in proteomics research field. In this paper, a dynamic time warping (DTW) model is introduced to find the optimal match between an experimental tandem mass spectrum (MS/MS) and a theoretical spectrum generated from a peptide sequence in a peptide database. Our approach can identify all types of PTMs in a blind mode. The test results on experimental spectra that contain no PTM, one PTM and two PTMs show that our approach can identify 95%, 93%, and 34% correct peptide sequences as the top 1 candidates, respectively, which is much better than other blind search methods.