Protein-expression profiling of serum is a common approach to the discovery of potential diagnostic and therapeutic markers of disease. Like any other proteome, the serum proteome is characterized by protein expression across a large dynamic range. This single facet requires the employment of fractionation procedures prior to detection of protein. The authors use a combination of conventional column chromatography with array-based chromatography to simplify the serum proteome into subproteomes, thus providing a greater representation of the serum proteome. Robotics is employed to increase the throughput of sample processing. These procedures result in large amounts of data that are analyzed through a series of preprocessing and postprocessing steps. A well-designed serum profiling project can therefore result in the discovery of statistically sound, clinically meaningful protein biomarkers.