Detection of Viral Proteins in Human Cells Lines by Xeno-Proteomics: Elimination of the Last Valid Excuse for Not Testing Every Cellular Proteome Dataset for Viral Proteins

Chernobrovkin, Alexey L.; Zubarev, Roman A.

In shotgun proteomics, proteolytic peptides are identified by comparing experimental tandem mass spectra of peptides with theoretical ones derived from the sequence databases [19] . Matching theoretical peptide to experimental spectrum has a probabilistic nature. Removing non-relevant sequences from the sequence database reduces the probability of random match between experimental and theoretical spectra, thus increasing the sensitivity of the method [20] . Thus, if there is no contamination, there is no reason to consider other species in the database search than the target one. However, choosing only one specific organism for matching the shotgun proteomics data amounts to assigning zero probability to the presence of other organisms in the sample. As has been demonstrated above, such an assumption can be dangerous. On the other hand, inclusion of several organisms in the database search amounts to assigning equal a priori probabilities to all proteins of all organisms, which reduces the search sensitivity and may lead to erroneous matches. Thus the simplest way to account for the possibility of viral presence, adding the sequences of all known viral proteins to the protein sequences database, is a faulty approach.