Use of cumulative distribution functions to characterize mass spectra of intact proteins

AbstractThe MH+ ions of matrix assisted laser desorption ionization time-of-flight (MALDI-TOF) spectra for a series of closely related but otherwise indistinguishable proteins were analyzed for singularity using a distribution free statistic, the Kolmogorov-Smirnov non-parametric statistic, K–S. The approach allows spectra which might otherwise be taken as identical, to be distinguished. Such analysis of the spectra may lead to a greater understanding of the chemistry of the proteins under investigation. The analysis requires only standard instrumentation. A standard data analysis protocol was developed and applied to generate a normalized cumulative distribution function (NCDF) for each spectrum. Differences in the NCDF for two different spectra were calculated and the maximum difference, Δmax compared to critical values of K–S. Values of Δmax exceeding the critical value of K–S are taken as the basis for rejecting the statistical null-hypothesis and assigning statistical significance to the differences in the two spectra. We have shown that this approach allows spectra of 1:1 mixtures of closely related recombinant proteins to be distinguished from either protein alone, and that mixtures of a 45 kDa protein and a labeled version of that protein can be distinguished from the pure material and from one another at the level of about 25%. In addition, we are able to use this approach to characterize the extent to which a synthetic glyococonjugation reaction has proceeded under circumstances of differing reaction times