protein group vs. subsumed entry

[Tomas]

I have a peculiar problem - two highly homologous proteins where I
identified bunch of peptides for one of them (including peptides
unique for it) and for the other I "identified" only peptides which
are shared by both proteins. The TPP (ProteinProphet) gives the first
one protein probability 1.0 and the second probability 0.0 (as
expected), but it separates them in the output, e.g. does not generate
a "protein group" where the first would be the main protein and the
other an "indistinguishable" group member. Instead what TPP did is
that it indicated the second one as a "subsumed entry" of the first
one as well as listing it separately as a protein with 0.0 protein
probability (giving it unique protein number).
Would anybody have explanation why not a group but subsumed entry and
what is the distinction between these two?

I use TPP v3.4 SQUALL rev.2, Build 200906181759.

Thanks,

Tomas

[Luis Mendoza]

Hello Tomas,

The behavior you describe is the default (or typical) for the original ProteinProphet tool.  As of TPP version 4.2.0. (released February 2009), subsumed proteins are "grouped" with their parent ones and displayed under a single entry.  Please upgrade your TPP to take advantage of this and other features, including the MININDEP option in ProteinProphet.  The release notes when this change was made can be found here:

http://groups.google.com/group/spctools-discuss/browse_thread/thread/a75b99f29c49ee98/845ace41d79afb83?lnk=gst&q=release#845ace41d79afb83

Hope this helps,

--Luis

--

You received this message because you are subscribed to the Google Groups "spctools-discuss" group.
To post to this group, send email to spctools...@googlegroups.com.
To unsubscribe from this group, send email to spctools-discu...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/spctools-discuss?hl=en.

[Tomas]

Luis,

we have deployed on our Linux cluster TPP version 4.3.2, but with that
I see identical result to the one I described before.
Is there issue with the Linux version?

Tomas

On Dec 17, 5:34 pm, Luis Mendoza <lmend...@systemsbiology.org> wrote:
> Hello Tomas,
> The behavior you describe is the default (or typical) for the original
> ProteinProphet tool.  As of TPP version 4.2.0. (released February 2009),
> subsumed proteins are "grouped" with their parent ones and displayed under a
> single entry.  Please upgrade your TPP to take advantage of this and other
> features, including the MININDEP option in ProteinProphet.  The release

> notes when this change was made can be found here:http://groups.google.com/group/spctools-discuss/browse_thread/thread/...
>
> Hope this helps,
> --Luis

>
> On Thu, Dec 17, 2009 at 2:46 PM, Tomas <tvai...@u.washington.edu> wrote:
> > I have a peculiar problem - two highly homologous proteins where I
> > identified bunch of peptides for one of them (including peptides
> > unique for it) and for the other I "identified" only peptides which
> > are shared by both proteins. The TPP (ProteinProphet) gives the first
> > one protein probability 1.0 and the second probability 0.0 (as
> > expected), but it separates them in the output, e.g. does not generate
> > a "protein group" where the first would be the main protein and the
> > other an "indistinguishable" group member. Instead what TPP did is
> > that it indicated the second one as a "subsumed entry" of the first
> > one as well as listing it separately as a protein with 0.0 protein
> > probability (giving it unique protein number).
> > Would anybody have explanation why not a group but subsumed entry and
> > what is the distinction between these two?
>
> > I use TPP v3.4 SQUALL rev.2, Build 200906181759.
>
> > Thanks,
>
> > Tomas
>
> > --
>
> > You received this message because you are subscribed to the Google Groups
> > "spctools-discuss" group.
> > To post to this group, send email to spctools...@googlegroups.com.
> > To unsubscribe from this group, send email to

> > spctools-discu...@googlegroups.com<spctools-discuss%2Bunsu...@googlegroups.com>

[Luis Mendoza]

Hello Tomas,

The latest version that we release is 4.3.1.; did you download the source code from our Sourceforge page?

You may also want to either delete the old stylesheet (say, the interact.prot.xsl file) or just choose to filter/sort the ProteinProphet results to make use of the new code (e.g. choose to Sort by Probability).  The reason for this is that the protXML viewer will use an existing stylesheet if it exists, which in this case was created by the old version of the tool.  (Note: you do not need to re-run the analysis.)

--Luis

To unsubscribe from this group, send email to spctools-discu...@googlegroups.com.