ABSTRACT
Proteomics is undergoing a rapid transformation from a qualitative global peptide sequencing discipline into a quantitative, reproducibility-driven practice. Nowhere is this more evident than in the rapidly expanding field of protein biomarker discovery where the general goal is to uncover statistically robust patterns of differential expression between or among subjects/samples representing distinct biological/temporal states. This report presents the analytical characterization of a label-free LC FT-ICR-MS workflow for differential proteomics analysis of human plasma. The key elements discussed include (i) methodologies for performing properly replicated experiments with highly reproducible sample preparation and analysis, including the use of internal standards to quantify variance at different steps in the process, (ii) a new methodology for performing sample re-analysis that uses off-line targeted robotic acquisition of complementary spectral data (e.g. ECD and/or IRMPD) to enhance the identification of differentially expressed peptides/proteins, and (iii) data processing pipelines capable of integrating the automatic statistical analysis of the label-free (LC-) MS signal, together with the intuitive and highly interactive curation and annotation of differential features using the output from standard sequence database search programs. We illustrate the application of the complete sample-to-annotated-differential-peptides (-proteins) workflow by describing the acquisition and analysis of a large multidimensional dataset from patients undergoing a controlled myocardial infarction resulting in an experimental setup in which each patients serve as their own control. Furthermore, we discuss a couple illustrative examples of mid-level proteins observed in this study whose plasma concentrations change consistently within and across patients, in a treatment- and time-dependent fashion.
