Poster #RP220
Prediction of Signature Peptides for Improved Protein Identification and Quantification
Kristofer Wårell*, Peter James*
*Protein Technology, Electrical Measurements, Lund, Sweden
A signature peptide is a uniquely defined part of a protein, which identify that protein in a larger set. Either peptide sequence or mass can be used as discriminating quality. Integrated in mass spectrometer software, signatures can be used to prioritise MS/MS candidates for improved identification, or guide selection of peptides with desirable characteristics such as antibody binding sites. Synthesised sequence signatures can improve current protein identification methods based on peptide fragmentation by using more mass spectrum data than just the peaks. Unique mass signatures can be synthesised with a label and added to protein digests prior to mass spectrometry. Comparison of spectrum peaks from synthesised peptides and their real counterparts facilitate protein quantification.
We have implemented a Perl program called SignPept to predict signature peptides. It performs theoretical tryptic digestion of proteins and ranks output based on sequence or mass of the resulting peptides. We have analysed all Swiss-Prot proteins from human and eight model organisms. For each species, more than 95% of all proteins are predicted to have at least one unique sequence signature. With a precision of 10 ppm Dalton, well over 90% of all proteins are predicted to have at least one unique mass signature.