Proteomic platforms that enable researchers to profile a high number of proteins across large sets of complex samples hold a great potential for biomarker discovery. LC-MS/MS-based methods can be used to analyse many samples without the need for protein labelling. As the analysis is a sequential process, the performance of the system has to be consistent throughout the entire experiment. In this study we used a set of spiked serum samples as well as a set of 55 clinical serum samples from schizophrenia patients and healthy volunteers to show that the label-free proteomic approach yields reproducible results across a large number of samples and can be used to accurately measure the relative protein abundance. Using this approach, we identified 1709 serum proteins covering a dynamic range of over three orders of magnitude. We believe that label-free quantitative proteomics is especially suited for biomarker discovery in large sample sets.