Extended Data Fig. 7: Protein quantification and chromatographic separation.

a, Comparison of GO term recovery and proteome coverage between SILAC ratios and iBAQ intensities from individual isotopologue channels in 20 SILAC datasets. Caption and error bars show the mean and standard deviation of the differences in the number of protein groups quantified and the AUC between SILAC ratios and iBAQ intensities. b, Recovery of proteins annotated to the same GO term in CF–MS experiments grouped by fractionation method. c, Difference in the median protein complex AUC between each pair of fractionation methods. Asterisks indicate pairs of measures of association with a p-value less than 0.05 in a two-sided Brunner-Munzel test. d, Regression coefficients for fractionation methods in multivariable statistical analysis, estimated by a linear model fit to the protein complex AUC and including terms for measures of association, missing value handling strategies, approaches to chromatogram normalization, and interactions between them. e, As in c, but for GO terms. f, As in d, but for GO terms. g, Recovery of known protein complexes in published CF–MS experiments grouped by fraction method, with each measure of association shown separately. h, As in g, but for proteins annotated to the same GO term. i, Recovery of individual protein complexes in published CF–MS experiments grouped by fractionation method. j, Number of protein complexes with at least three subunits detected exclusively by one separation method, left, and resolved significantly better by one of the four separation methods, right, across 67 human and mouse CF–MS datasets. k, Examples of protein complexes resolved best by each of the four separation methods. Inset text shows the median AUC.