Optimal computational comparison of mass spectrometric peptide profiles of alternative hydrolysates from the same starting material

•Computational approaches were applied to food hydrolysate peptide profiles to evaluate relative similarities and differences.•Terminal amino acid abundances provide optimal grouping of peptide profiles.•The grouping method requires no a priori knowledge of cleavage patterns.

Food protein hydrolysates are complex mixtures that are increasingly being analysed by tandem mass spectrometry. Given a single starting material, many alternative peptide profiles are achievable under varying hydrolysis conditions. To date, characterisation of the relative similarities and differences between such peptide profiles remains largely unstudied. Here, we investigate optimal computational methods for grouping peptide profiles of hydrolysates derived from the same starting material. Using an experimental bovine milk dataset, we evaluated how these methods grouped either technical replicates, or distinct samples with known cleavage profiles. Analyses performed using only the presence and abundance of peptides were found to be suboptimal for achieving effective sample grouping. Using the amino acid distribution at both termini of peptides was more efficient at grouping replicate samples; however, this approach lacked suitable discrimination between distinct samples. By extending the termini approach to incorporate the abundance associated with terminal amino acids, optimal grouping was achieved. We therefore suggest that grouping peptide profiles of hydrolysates from the same starting material should rely on a combination of N and C terminal amino acid frequency and abundance. Importantly, this approach requires no a priori knowledge of enzyme specificities, making it generally applicable to diverse sets of food matrices.