It is completely normal. The fact is that iProphet is a better classifier, which means that it is removing false matches while retaining the good matches at a higher rate than PeptideProphet. Intuitively, to achieve the same FDR with iProphet you have to dig deeper in the data than you do with PeptideProphet to find the false matches that make up your "false discoveries". Also, iProphet probabilities are probabilities corresponding to the entire group of peptides that all share the same sequence (distinct peptide sequence level), so iProphet probabilities may seem too conservative when evaluated on the PSM level. iProphet probabilities/FDRs are best evaluated on distinct peptides, where they should be similar in accuracy to PeptideProphet probabilities/FDRs on the PSM level.
-David On Mon, Oct 27, 2014 at 6:58 AM, Huoming Zhang <[email protected]> wrote: > Dear TPP team, > > I used MAYU for filtering protein identification. When I perform MAYU on > iprophet result, it shows that the IP/PPs is jumping for 0.99914 to 0.6976 > (number in red), which seems not normal; The MAYU for peptideprophet result > shows more reasonable (number in blue, no big jump for the IP/PPs); what > could be the reason for this big difference? can I use MAYU on > peptideprophet results? Thanks. > > mFDR IP/PPs TP_protID 0 *0.99914* 2484 0.001 *0.697643* 6770 0.002 > 0.504006 7085 0.003 0.389325 7259 0.004 0.289737 7430 0.005 0.236973 > 7472 0.006 0.197837 7509 0.007 0.162701 7578 0.008 0.135955 7630 0.009 > 0.11707 7677 0.01 0.098871 7750 > > mFDR IP/PPs TP_protID 0 *0.9997* 1747 0.001 *0.9728* 6615 0.002 0.949 > 6900 0.003 0.9245 7094 0.004 0.8969 7235 0.005 0.8642 7341 0.006 > 0.8347 7374 0.007 0.8077 7403 0.008 0.7719 7461 0.009 0.7316 7539 0.01 > 0.6991 7580 > > -- > 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 http://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 http://groups.google.com/group/spctools-discuss. For more options, visit https://groups.google.com/d/optout.
