Hi John, Yes, the semi-supervised and semi-paramteric features from Alexey's papers are implemented. Specifying a decoy tag to PeptideProphet makes it semi-supervised and selecting the non-parametric model makes it semi-parametric; the one remaining parameter in the distributions is the apportionment of non-decoy hits among the positive and negative distributions (which in this mode do not follow the shape of a predetermined parametric distribution e.g. Gamma for negatives and Gaussian for positives, but are learned using Kernel Density Estimates).
-David On Tue, Sep 21, 2010 at 7:31 AM, John Damask <[email protected]> wrote: > Hi David, > Thanks for the response. The release notes are helpful. And no, I > haven't noticed increased decoy hits w/ higher prob...but then again I > haven't really looked. > > In addition to the release notes, is there any documentation about > which features from Alexey's papers have made it into the prophets? > For example, Choi, Gosh, & Nesvizhskii 2007 JPR discuss the benefits > of "variable component mixture model and semiparametric mixture > model". I would guess one of these models is invoked if I check the > "Use non-parametric model" option but I'm just not sure. To be clear, > I'm not asking for full compliance with the literature, I just want to > know if PeptideProphet differs from it. > > Thanks, > John > > On Sep 20, 1:24 pm, David Shteynberg <[email protected]> > wrote: >> Hi John, >> >> The target is for the quality of the software to improve with >> subsequent releases. Do you notice that you have more incorrect >> (decoy) hits with higher probability? BTW, the latest version is 4.4 >> not 4.3. For more information you can check the release >> noteshttp://tools.proteomecenter.org/wiki/index.php?search=release+notes&g... >> >> -David >> >> On Tue, Sep 14, 2010 at 7:45 AM, John Damask <[email protected]> wrote: >> > Hi, >> > We recently upgraded an old version of TPP (Squall v3.3) to v4.3 rev1. >> > We've noticed an increase in the number of high confidence (<1% FDR) >> > peptides and proteins. Digging into the pepXml files we've seen cases >> > where, for example, a peptide previously assigned a probability of < . >> > 5 now scores > .9. Note that the Mascot scores haven't changed >> >> > Where can I learn about the differences between versions of the >> > Prophets? >> > Thanks >> > John >> >> > -- >> > You received this message because you are subscribed to the Google Groups >> > "spctools-discuss" group. >> > To post to this group, send email to [email protected]. >> > To unsubscribe from this group, send email to >> > [email protected]. >> > For more options, visit this group >> > athttp://groups.google.com/group/spctools-discuss?hl=en. > > -- > You received this message because you are subscribed to the Google Groups > "spctools-discuss" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/spctools-discuss?hl=en. > > -- You received this message because you are subscribed to the Google Groups "spctools-discuss" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/spctools-discuss?hl=en.
