View article

To the editor: Mass spectrometry (MS)-based proteomics is performed in different experimental formats, which today almost always involve matching fragmented peptides to a database1. Different protein sequence databases strike different balances between inclusion of as many protein isoforms as possible and minimal redundancy. The European Bioinformatics Institute (EBI) International Protein Sequence (IPI) database2 contains about 68,000 human protein entries and is widely used in proteomics. IPI stores the longest sequence available for each entry, which means that the N-terminal peptide of secreted proteins, for example, appears to be ‘half-tryptic’and is not reported when requiring complete sequence specificity of the protease used to degrade proteins to peptides3. IPI also does not contain information about conflicting sequence entries contained in the Swiss-Prot database. Furthermore, the genomes of …