Hello Rachel, The table shows the model-based error estimate on the PSM level, which tends to be conservative in the iProphet analysis. If you take the maximum iProphet probability over any peptide sequence and compute the peptide level error rate these should be closer to the decoy estimates. With the conservative estimate provided by iProphet you should count fewer decoy hits than predicted incorrect IDs by the model. Also, the decoys are not the only source of incorrect only the true negatives, there is also a population of incorrect hits to targets that you can estimate based on the rate of decoys in the database.
I hope this helps. -David On Tue, Jun 26, 2018 at 12:17 PM, Rachel Nadeau <[email protected]> wrote: > Hello, > > I have searched a dataset using Comet (target-decoys search) and I am now > trying to filter it using iProphet. I used the error table to decide on a > probability threshold based on the Error_Rate. However, in the exported > data the number of decoy hits above the probability threshold does not > match the number of incorrect matches (num_incorrect) in the error table. > Should those be matching or am I missing something? > > Please find enclosed the link for the Models file (containing the error > table), the iProphet xml output file, and the extracted iProphet output > file above a probability cut-off of 0.7627, which should correspond to 917 > incorrect identifications. > https://drive.google.com/open?id=1zFyrJCaK2ZdpfJ-1VmiIvv7S38yZSZSD > > Thank you very much for your help! > > Sincerely, > > Rachel > > -- > You received this message because you are subscribed to the Google Groups > "spctools-discuss" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at https://groups.google.com/group/spctools-discuss. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "spctools-discuss" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/spctools-discuss. For more options, visit https://groups.google.com/d/optout.
