<https://lh3.googleusercontent.com/-14WvTOiLMM4/VCqmP8FV3oI/AAAAAAAAA6U/Rc4JO7qFmzQ/s1600/modelproblem.png> Hi all, we recently had some problems with PeptideProphet runs where the model failed for 2+ ions. Attached is a summary of replicate measurements of the same biological sample. The pep3d maps look almost identical (apart from the colored probabilities of course) and also the discriminant score distributions are very similar. However, in case of replicate B all was fine, while for replicate A PeptideProphet is unable to model the 2+ peptides. All I found in the command output is "WARNING: mixture model quality test failed for charge 2+". When I run PeptideProphet in the semi-parametric mode (i.e. reveal the decoy label using the -d option) the model works just fine... But usually we want to keep the decoy hits as a second level of FDR estimation.
Anybody experienced something similar, or has ideas why the "model quality test" fails and how to improve it? thanks a lot, Ulrich -- 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.
