Figure 2.

The number of proteins not supported by the best scoring peptide inference can be estimated by a decoy model. Here, we plotted the reported fraction of proteins with incorrect best scoring peptide inference as a function of the classical decoy–target ratio, for ten randomized simulations for three different values of the absent protein fraction, ${\pi }_A$, and 20 000 peptide inferences. The classical decoy–target ratio accurately matches the fraction of proteins with incorrect best scoring peptide inference.