Publication at Faculty of Mathematics and Physics |
2010
Abstract
Tandem mass spectrometry is a widely used method for protein and peptide sequences identification. Since the mass spectra contain up to 80% of noise and many other inaccuracies, there still exists a need for more accurate algorithms for mass spectra interpretation.
The sizes of protein databases grow rapidly and the methods for indexing these databases in order to interpret mass spectra become very popular. The parametrised Hausdorff distance, suitable for non-metric search, is presented in this paper.
It models the similarity among tandem mass spectra very well and it is able to match the spectrum to correct peptide sequence in many cases without any post-processing scoring system.