Label-free protein quantitation by spectral counting has proved to be a useful method for accurate protein profiling. However, much work still remains to be done in applying useful and valid statistical methods for validation of shotgun results. Due to the large number of protein identifications involved, any t-test will invariably yield false positives; this is known as the multiple testing problem. My work focuses on the development of a novel form of statistical validation based on control-control experiments, where the statistical parameters for empirically minimising false discovery rates are determined from triplicate analyses of biologically identical material. An appropriate statistical measure (q value) is then carried forward into the analysis of control vs different samples. Through this additional statistical step, the resulting output comprises highly confident protein identifications with a minimised protein quantitation false discovery rate (PQ-FDR), which represents a work-around to solve the multiple testing problem.