Doubling down on peptide phosphorylation as a variable mass modification

Authors: Cooper, Bret

Submitted to: Proteomics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/18/2016
Publication Date: 9/1/2016
Publication URL: http://handle.nal.usda.gov/10113/5852180
Citation: Cooper, B. 2016. Doubling down on peptide phosphorylation as a variable mass modification. Proteomics. 16:2444-2447. Doi 10.1002/pmic.201500440.

Interpretive Summary: Mass spectrometry is used to identify proteins and peptides in biological samples. The mass spectrometer produces data that is interpreted by software and the software assigns a score indicating a particular peptide was identified. The purpose of this research report is to invalidate a common belief that searching for phosphorylation, a protein modification, lowers the confidence in identification scores. We have used a soybean leaf data and human cell data to prove our hypothesis. Debating data confidence is a common, natural part of scientific discovery and leads to new experiments and findings that bolster or overturn claims. Therefore, this report is most likely to influence scientists at universities, government agencies and companies who want to assign better confidence to the peptides and proteins they discover by mass spectrometry.

Technical Abstract: Some mass spectrometrists believe that searching for variable post-translational modifications like phosphorylation of serine or threonine when using database-search algorithms to interpret peptide tandem mass spectra will increase false positive rates. The basis for this is the premise that the algorithm compares a spectrum to both a non-phosphorylated peptide candidate and a phosphorylated candidate, which is double the number of candidates compared to a search with no possible phosphorylation. Hence, if the search space doubles, the false positive rate could increase accordingly as the algorithm considered more candidates to which false matches could be made. In this study, it is shown that the search for variable phosphoserine and phosphothreonine modifications does not always double the search space or unduly impinge upon the false discovery rate. A breakdown of how one popular database-search algorithm deals with variable phosphorylation is presented.