Evaluating the probability that ms/ms spectral data matches candidate sequence data

Abstract

In one aspect of the present invention a new database search methodology is provided that provides a probability that spectral data from a non-ergodic reaction via mass spectrometry matches a candidate sequence from a set of sequences in a database by random. The methodology comprises two parts. The first part pre-processes the spectral data and retains only the most relevant data for the database search. The second part comprises searching a database using the pre-processed spectrum to assign a probability or expectation that the spectrum matches a candidate sequence from a set of sequences in a database by random. The search methodology uses a new probability model, a compound distribution based on the number of product ion mass-to-charge ratios and the number of intensity values that are shared between the product ion spectral data and the sequence database, to accurately predict the probability of the peptide identification being a correct match, and not a random event.