| While genomic techniques are well-developed, successful technologies, similar techniques to analyze the protein component of a cell are still emerging. This fledgling field of proteomics is rapidly developing and closing the last gap in direct, high-throughput study of all phases of the central dogma. Although new, sensitive, and highly accurate instrumentation have revolutionized the use of liquid chromatography - tandem mass spectrometry (LC-MS/MS) for proteomic investigation, it is only in conjunction with robust computational methods that the power of large-scale studies can be truly harnessed.;This dissertation addresses the development of these computational tools to understand all aspects of protein regulation. The biological activity of any protein is chiefly controlled by its primary amino acid sequence, its abundance, and its modification state and mass-spectrometry-based methods are especially suited to study of these three properties in a massively parallel fashion. Here, I address each of them in turn.;Protein identification is arguably the most successful application of LC-MS/MS, and has been widely employed in a diverse array of experiments. However, these methods are limited in their depth of analysis of the proteome, and it is crucial that LC-MS/MS methods employ the right experimental approach to maximize their reach. I present a critical evaluation of how experimental design, methods construction, and data analysis affect these results, both in complex mixtures as well as in the phosphoproteome.;Beyond identification, the most exciting application of mass spectrometry is to a global understanding of protein abundance. Nevertheless, there are significant challenges to gathering quantitative data from LC-MS/MS. In particular, low signal levels within the mass spectrometer are thought to harm quantitative accuracy. Here, I verify this hypothesis and present a solution based upon accurate mass, reducing the dependency of accuracy upon signal-to-noise. I also address the uncoupling of quantitative study from protein identification, improving the yield of large-scale, quantitative experiments.;Lastly, the application of LC-MS/MS to elucidate post-translational modifications (PTMs) is a critical area of research. While current methods can assign PTMs, they require a priori knowledge of the modification mass. I present progress toward an algorithm which identifies and localizes PTMs without such prior knowledge. |
