Have you tried the recently released 4.8? Seems like it should be working on this data, but perhaps 4.7.0 has an issue. Another parameter to play with is described here:
http://goo.gl/lfBGjK -David On Mon, Dec 1, 2014 at 7:36 AM, Ulrich Omasits <ulrich.omas...@gmail.com> wrote: > Dear TPP developers & users, > we have some serious trouble with PeptideProphet (TPP v.4.7.0) on samples > from our QExactive: > > For rather complex samples (yeast lysate in this case), PeptideProphet often > fails to model the 2+ peptides (sometimes also 4+) despite a very obvious > bimodal xcorr and fval distribution. We do a decoy search, so I could use > the decoys in PeptideProphet to build the models in a semi-parametric or > non-parametric way. However, I would prefer to use the decoy hits as FDR > validation. > > Using standard xinteract parameters (-p0 -OApl) the 2+ model fails. When > forcing the model (-p0 -OAplF) a model will be displayed, but still all the > 2+ peptides in the pep.xml have a probability of zero!! > Using the semi-parametric model (-p0 -OApld -dDECOY) the 2+ model works, but > does not look very convincing to me (see also attached graphs). > Using the non-parametric model (-p0 -OApldP -dDECOY) fails, but when forcing > it (-p0 -OApldPF -dDECOY) the displayed model looks really nice, but again > all the 2+ peptides in the pep.xml have a probability of zero!! > (see also attached graphs) > > > So I have two major questions: > > 1) How can I get the peptide probabilities from a forced model (using the > -OF option)? > > 2) Could it be that the model-quality checks do not really work for the high > identification rate of fast instruments like the Qexactive, where the > negative distribution appears as just a shoulder on the left of the larger > positive distribution? > > Has anybody any ideas what else I could try? I could also share the source > files for this run. > > Thanks a lot! > Ulrich Omasits > Institute of Molecular Systems Biology > ETH Zürich > > -- > 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 spctools-discuss+unsubscr...@googlegroups.com. > To post to this group, send email to spctools-discuss@googlegroups.com. > 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 spctools-discuss+unsubscr...@googlegroups.com. To post to this group, send email to spctools-discuss@googlegroups.com. Visit this group at http://groups.google.com/group/spctools-discuss. For more options, visit https://groups.google.com/d/optout.
