A method for matching proteins in databases with unknown samples using data derived from peptide product ion masses is described. The power of the method is due to the speed of analysis and to the ability to locate proteins in a database even when the experimental data show anomalies due to derivatization or post-translational modification.