The accuracy of quantification obtained using the iTRAQ labelling methodology for measuring protein ratios more extreme than 1:1 was investigated. A comparison of nLC-ESI MSMS and nLC-MALDI MSMS analysis routes was performed. A fixed concentration of a standard six protein mix was spiked with two proteins at a range of concentrations. The two data analysis programmes, Mascot and ProteinPilot Paragon, were also compared. Whilst the lower ratios could be measured accurately, greater discrepancies were seen for the higher ratios, particularly by nLC-ESI MSMS. Filtering out the weaker reporter ion signals improved the accuracy of the ratios: this is likely due to several factors which are explored in more detail. Overall, analysis by nLC-MALDI MSMS followed by Mascot interpretation gave the most accurate results.