Proteome analysis by tandem mass spectrometry: Improvements and biological applications.

摘要

Protein identification by liquid chromatography and tandem mass spectrometry (LC-MS/MS) has evolved as the dominant tool for performing large scale proteome analyses. The recent success of this method can be attributed to improved speed and accuracy of mass spectrometers, and computational solutions to the problem of managing vast amounts of data produced in mass spectrometry experiments. This dissertation presents three such computational solutions: (1) a 'target/decoy' searching approach for effective and efficient false positive rate estimations from large sets of mass spectra; (2) a model of the processes that contribute to peptide fragmentation in the mass spectrometer for harnessing measured fragment ion intensities to more sensitively select correctly interpreted mass spectra; and (3) a method for combining complementary analysis techniques, which greatly enhances large mass spectrometry-based analyses. Building on this foundation of ways to generate large, high-confidence proteomic data sets, I present progress in an ongoing experiment that so far, has measured the degradation rates of 1,800 proteins---far more than have ever been measured in a single experiment. These measurements provide novel evidence suggesting numerous regulators or diverse cellular processes.