Fig. 3: IceR enables robustly improved sensitivity for wide ranges of typically varied MS analysis parameters.

a Effect of TopN, gradient length and sample injection amount on numbers of identified E. coli proteins with at least two quantification events (top panel), missing values (centre panel) and CVs of quantification (bottom panel) in MaxQuant (grey) and IceR (orange) data. Each condition was analysed in n = 3 independent samples. centre line of boxplots, median; box limits, upper and lower quartiles; whiskers, 1.5× interquartile range. b As in a but showing the effect of parameters on sensitivity (extracted from ROC curves at 95% specificity, upper panel), detection of true positives (center panel) and detection of false positives (bottom panel) over all pairwise differential expression analyses in MaxQuant (grey) and IceR (orange) data. Each condition was analysed in n = 3 independent samples. Asterisks indicate significance level which was determined by one-sided t-tests based on log-transformed counts: *p-value < 0.05, **p-value < 0.01.