Large-scale Protein Identification In Tandem Mass Spectrometry Data Evaluation

It can be obtained tens or hundreds of thousands of MS/MS spectra from the identifications of complex system of proteins, as the difference of sample' s complexity and the strategies for protein identification. Ten percent to twenty percent of these spectra can be precise identified, but the others are poor in quality, this makes that more false positive peptides were identified by random assignment, and there are more overlapping between the positive identifications and false positive identifications, then these lead to more false positive protein identifications. The literatures for large-scale protein identifications have more or less false identification results in it, even some labs with little experience on biomass think that all the protein identifications based on mass spectrometry are correct and don' t give the in-depth analysis on them.The False positive rate of protein identifications for tandem mass spectrometry is reduced and the reliability of these data is improved and the data come from different strategies, mass spectrometry and labs can be compared by using the method of random database searching. We apply this method on the analysis of the data from human liver primarily in this study. Setting the same cutoff for shorter and longer peptides is obviously inappropriate in high-throughput protein identifications based on Sequest because the Xcorr value is dependent on assigned peptide size. The Xcorr value was normalized , then True positive peptide identifications increased by 13. 5%, 17.5% and 13.6% for singly charged peptides, doubly charged peptides and triply charged peptides for a mixture of 18 known proteins, respectively, in total, an increase of 15. 8% was obtained for all charged peptides, when normalized Xcorr threshold and DeltaCn threshold were used. We suggest that application of random database searching strategy with normalized Xcorr can improve peptide identifications and reduce false positive rate and false negative rate of peptide identifications in shotgun proteomics.