On the Statistical Significance of Compressed Ratios in Isobaric Labeling: A Cross-Platform Comparison

Isobaric labeling is gaining popularity in proteomics due to its multiplexing capacity. However, copeptide fragmentation introduces a bias that undermines its accuracy. Several strategies have been shown to partially and, in some cases, completely solve this issue. However, it is still not clear how ratio compression affects the ability to identify a protein's change of abundance as statistically significant. Here, by using the "two proteomes" approach (E. coli lysates with fixed 2.5 ratios in the presence or absence of human lysates acting as the background interference) and manipulating isolation width values, we were able to model isobaric data with different levels of accuracy and precision in three types of mass spectrometers: LTQ Orbitrap Velos, Impact, and Q Exactive. We determined the influence of these variables on the statistical significance of the distorted ratios and compared them to the ratios measured without impurities. Our results confirm previous findings1-4 regarding the importance of optimizing acquisition parameters in each instrument in order to minimize interference without compromising precision and identification. We also show that, under these experimental conditions, the inclusion of a second replicate increases statistical sensitivity 2-3-fold and counterbalances to a large extent the issue of ratio compression.