Influence of various endogenous and artefact modifications on large‐scale proteomics analysis

RATIONALE: Some large-scale proteomics studies in which strong cation exchange chromatography has been applied are used to determine proteomes and post-translational modification dynamics. Although such datasets favour the characterisation of thousands of modified peptides, e.g., phosphorylated and N-α-acetylated, a large fraction of the acquired spectra remain unexplained by standard proteomics approaches. Thus, advanced data processing allows characterisation of a significant part of these unassigned spectra. METHODS: Our recent investigation of the N-α-acetylation status of plant proteins gave a dataset of choice to investigate further the in-depth characterisation of peptide modifications using Mascot tools associated with relevant validation processes. Such an approach allows to target frequently occurring modifications such as methionine oxidation, phosphorylation or N-α-acetylation, but also the less usual peptide cationisation. Finally, this dataset offers the unique opportunity to determine the overall influence of some of these modifications on the identification score. RESULTS: Although methionine oxidation has no influence and tends to favour the characterisation of protein N-terminal peptides, peptide alkalinisation shows an adverse effect on peptide average score. Nevertheless, peptide cationisation appears to favour the characterisation of protein C-terminal peptides with a limited to no direct influence on the identification score. Unexpectedly, our investigation reveals the unfortunate combination of the molecular weight of N-α-acetylation and potassium cation that mimics the mass increment of a phosphorylation group. CONCLUSIONS: Since these characterisations rely upon computational treatment associated with statistical validation approaches such as 'False discovery rates' calculation or post-translational modification position validation, our investigation highlights the limitation of such treatment which is biased by the initial searched hypotheses.