Hi Allen,
May I know what software are you using? I'd like to find a similar
EM-like PeptideProphet algorithm written in other languages. Thanks.
On Monday, September 10, 2007 at 11:28:52 PM UTC-7, Alden wrote:
>
> Dear professor :
>
> Thanks for your reply : )
> Yes, I am using a different software which have the same EM-like
> algorithm with ProteinProphet.
> Because the examples showed here is simple, the protein entries
> probability can be proved to be correct through manual calculation
> according to the EM-like algorithm.
> Still, I am not sure whether the peptides' probability can be
> changed or not using the nsp adjustment if their initial probability
> are set as 1.00
> Thanks!
>
>
> On 9月11日, 上午4时45分, "[email protected]" <[email protected]> wrote:
> > Dear Alden,
> >
> > Do I understand correctly that you are using not ProteinProphet but a
> > different software? If so, it would be hard to say what wring with
> > your program. Maybe not enough iterations, or the weights are
> > calculated incorrectly.
> >
> > Alexey Nesvizhskii
> > University of Michigan
> >
> > On Sep 8, 12:02 pm, Alden <[email protected]> wrote:
> >
> > > Hello everyone!
> > > I have noticed that in the tutorial about protein inference
> > > problem [ Molecular & Cellular Proteomics, 2005. 4(10): p.
> > > 1419-1440.], the authors said that "all subset and subsumable protein
> > > entries are assigned zero probability" using ProteinProphet.
> > > When I calculate protein probability with my own software which
> > > have the same EM-like algorithm with ProteinProphet, it seems that
> > > that the subset & subsumable protein entries are not always assigned
> > > zero probability.
> > > Take the following illustration as an example :
> >
> > > Pep1 Pep2 Pep3
> > > Prot A - - -
> > > Prot B - -
> > > Prot C - -
> >
> > > Prot B & Prot C are subset of Prot A. We assume the initial
> > > probability of peptide 1, 2 and 3 are 1.00, so the adjustment of
> > > peptide probability for NSP doesn't work according to eq (5) [ Anal.
> > > Chem. 2003, 75, 4646-4658]. Finally,using the EM-like algorithm, the
> > > probability of each protein is,
> > > Prot A: 0.93
> > > Prot B: 0.54
> > > Prot C: 0.54
> > > The subset protein entries' probability is not zero !
> >
> > > Another simple example shows that subsumble protein's probability
> > > is assigned nonzero. Look at the following illustration:
> >
> > > Pep1' Pep2' Pep3' Pep4'
> > > Prot A' - -
> > > Prot B' - -
> > > Prot C' - -
> >
> > > Prot B' belongs to subsumble protein and the probability is
> > > computed as 0.62 (we also assume the initial probability of peptides
> > > are 1.00).
> >
> > > Can anyone explain why this happens?
> > > Looking forward to your reply :-)
>
>
--
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.
