Files
Abstract
Mass spectrometry based proteomics has become one of the most powerful tools used to determine protein structure, function, and expression. The recent rapid expansion of the field is a result of significant improvements in the throughput of mass spectrometers and the availability of affordable and powerful computers. Due to the improvements to mass spectrometers, LC-MS/MS experiments are able to generate tens of thousands of MS/MS spectra making automated interpretation a necessity. The results of automated database searches must be scrutinized very carefully.In this study, the incidence of false positives from search results was explicitly measured by using a database annotated with randomly generated proteins. True positives were measured as the identification of a subset of individually validated recombinant proteins that were spiked intothe experimental sample. Searches were performed under a variety of high and low mass accuracy settings, and the performance of a few methods of statistical validation was analyzed.This investigation shows that post-search statistical validation should be a mandatory step in a shotgun proteomics experiment. Furthermore, high resolution mass accuracy and resolution ion fragment scanning should be used as soon as mass spectrometers can operate in those modes without significant losses to scan times and throughput.