Figure 4:

Calibration of the residue evidence score via dynamic programming. (A) The figure plots the p-value, as calculated via dynamic programming, versus the rank p-value, for decoy PSMs from a Plasmodium dataset. The lines y = x (solid line), y = 2x (dotted line) and y = 0.5x (dotted line) are included for reference. (B) The figure plots, for the Plasmodium dataset, the number of PSMs accepted as a function of q-value threshold, for four different database search methods: XCorr calibrated via dynamic programming using low-resolution m/z bins, uncalibrated XCorr using high-resolution m/z bins, uncalibrated res-ev, and res-ev calibrated via dynamic programming. Note that a series corresponding to calibrated XCorr using high-resolution m/z bins is not included, because dynamic programming cannot be carried out in conjunction with small m/z bins, as explained in the Introduction.