When performing MBR, we attempt to locate a decoy peak for each acceptor peak. Acceptor peaks are located by predicting a retention time and looking for the theoretical peptide mass at that retention time. Decoy peaks are located by choosing a "random" retention time and looking for the theoretical peptide mass at the random retention time.
PEP Calculation for MBR Peak
All MBR peaks are fed into a binary classifier (similar to the one used within MetaMorpheus) and semi-supervised learning is performed to calculate a PEP for each peak.
MBR Q-Values
For each file, Q-Values are calculated for each MBR peak by ranking them in PEP order and performing target-decoy competition
Major changes to MBR within FlashLFQ.
Peak Decoys
When performing MBR, we attempt to locate a decoy peak for each acceptor peak. Acceptor peaks are located by predicting a retention time and looking for the theoretical peptide mass at that retention time. Decoy peaks are located by choosing a "random" retention time and looking for the theoretical peptide mass at the random retention time.
PEP Calculation for MBR Peak
All MBR peaks are fed into a binary classifier (similar to the one used within MetaMorpheus) and semi-supervised learning is performed to calculate a PEP for each peak.
MBR Q-Values
For each file, Q-Values are calculated for each MBR peak by ranking them in PEP order and performing target-decoy competition