Search

Normalisation of data to minimise the impact of technical variation on comparative sample analysis is often carried out. Using SELDI data as a model, we have examined the effects of normalisation by TIC which is commonly used for MS data. Significant intergroup differences in normalisation factor were found for serum profiles which could not be explained by experimental factors, implying that normalisation by TIC may in some situations also normalise biological differences and should be systematically evaluated.

A database of 5448 peptide tandem mass spectra acquired in a quadrupole time-of-flight mass spectrometer was generated for peptides derived from proteins digested with trypsin. Peptides were identified from their mass spectra by the Mascot algorithm. Statistical models were then used to investigate factors influencing the abundance of ions formed. Separate models were formulated for b and y ions as it was thought that different factors may influence the formation of each type of ion. Several factors were found to have a highly significant influence on the abundance of ions formed. These include the actual mass of the ion formed after fragmentation as well as the location of the cleavage. The composition of the fragmenting peptide was also found to be important, and amino acids either side of the fragmentation site influenced the abundance of ions produced. To increase understanding of fragmentation mechanisms, the effect of several physicochemical properties of these residues was also investigated in a separate model. In conclusion, the models formulated for b and y ions provide useful characterization of the abundance of ions formed, and this information could be used to develop improved algorithms for peptide identification.