Not updating Settings file

[GoogleCodeExporter]

Not updating Settings file

[GoogleCodeExporter]

Thanks for reporting your DeNovoGUI issue. And thanks for the nice comment 
about our software!

You are indeed correct in that the number of solutions seems to be reset in the 
current online release. A new version of DeNovoGUI is just around the corner 
though. If everything goes well and we don't come across any new issues it 
should be out sometime next week.

In addition to fixing the number of solutions issue the major update is that it 
will now also support DirecTag in addition to PepNovo+. :)

Original comment by [email protected] on 14 Mar 2014 at 3:34

• Changed state: Started

[GoogleCodeExporter]

Harold, great to hear about the new release. Please keep in mind the 
possibility of exporting n top peptide hits per spectrum to BLAST export. This 
would be very useful for large files which take long Blasting times.  Once you 
set a threshold score, you dont necessarily want ALL the hits per spectrum that 
pass that threshold, since Blast will still select the most overlapping 
sequence... 
Another nice addition that Lutefisk had is that in the case of c- and n- gaps, 
if the mass would match the combinatorial of x amino acids, they would appear 
in brackets, meaning that their order was unconfirmed... This sometimes helps 
peptide confirmation according to the blasted sequence.. It often occurs that 
the gap agrees with the DB residues to the next enzymatic cleavage site.

Hope this helps, and again, thanks for your freeware collaboration to the 
proteomic community!

Alex

Original comment by [email protected] on 14 Mar 2014 at 8:07

[GoogleCodeExporter]

Thanks for the ideas for new features. The "export n top peptide hits per 
spectrum to BLAST export" should not be a problem. I'll include it in the 
upcoming release.

Showing the combinatorial of x amino acids in the case of gaps also sounds like 
a good idea. I'll look into how difficult this would be to implement.

We are also working on getting rid of the need for BLAST altogether by mapping 
the best hits to a user provided FASTA database file ourselves. Will most 
likely not be in this next release, but should not be too far off.

BTW, if you like DeNovoGUI, be sure to also check out SearchGUI and 
PeptideShaker (http://searchgui.googlecode.com and 
http://peptide-shaker.googlecode.com). For more details on the use of our tools 
see also our tutorials:  http://compomics.com/bioinformatics-for-proteomics.

Original comment by [email protected] on 14 Mar 2014 at 11:09

[GoogleCodeExporter]

Nice... the fasta search would be a nice addition. Lutefisk came bundled
with a similar program called CIDentify (
http://www.hairyfatguy.com/lutefisk/).

As for SearchGUI and Shaker, we recently got a VelosPro Orbitrap with
software (ProteomeDiscover) support for 5 years... so I'll stick to that.
Ocasionally I use other DB search engines for very particular stuff. When
support is over, I'll look into alternative opensource programs.

Nice weekend!

Alex

PS: can you send me an email when the update is ready?

Original comment by [email protected] on 14 Mar 2014 at 11:55

[GoogleCodeExporter]

Yes, I will let you know when the new DeNovoGUI version is available for 
download.

Original comment by [email protected] on 15 Mar 2014 at 12:01

[GoogleCodeExporter]

DeNovoGUI 1.3.0 has now been released. It fixes the problem with the number of 
solutions not being changed and also adds the requested "export the Nth most 
highest match for each peptide" when exporting to BLAST. 

DeNovoGUI 1.3.0 adds support for DirecTag and tag to FASTA file mapping.

For the complete list of changes please see: 
http://code.google.com/p/denovogui/wiki/ReleaseNotes.

Best regards,
Harald

Original comment by [email protected] on 26 Mar 2014 at 9:16

• Changed state: Fixed

[GoogleCodeExporter]

Harald,

Thanks for the update... just to note that for some reason not ALL the
unimod modifications are available (first time I try this, could have been
also there in previous version of denovogui). As an example, Unimod's
Accesion #64 (PSI-MS name: Succinyl) is not there, and I had to manually
add it, however I would have expected it to be there if DenovoGui is
reading Unimod's full list of modifications, right??

I'm trying 1.3 now... I'll let you know how things are working out!

Original comment by [email protected] on 27 Mar 2014 at 3:15

[GoogleCodeExporter]

DeNovoGUI does not actually access Unimod directly (or indirectly). Due to the 
way the PTMs are handled in SearchGUI to support all the search engines we have 
to map them to a common (OMSSA based) format in the backend. But as you already 
noticed you can quite easily add your own PTMs.

Harald

Original comment by [email protected] on 27 Mar 2014 at 3:23

[GoogleCodeExporter]

Thanks Harald... yes indeed very user friendly to add the modifications.

Could give me your own (quick) thoughts on which particular scenario
DirectTag would provide information Pepnovo wouldn't? I haven't used
DirecTag before, just read the paper, and wondering how much Pepnovo and
DircetTag overlap... your thoughts?

Alex

Original comment by [email protected] on 27 Mar 2014 at 3:33

[GoogleCodeExporter]

Well, I cannot say that there are clear cases where the one would do better 
than the other. But as they are different algorithms the results will not be 
identical. So hopefully this will result in the total amount of useful tags 
increasing.

Note also that DirecTag only creates short(er) tags, while PepNovo+ tries to 
sequences as much of the spectrum as possible. One consequence of this is that 
DirecTag is a lot faster than PepNovo+.

The main reason for adding DirecTag though was that the PepNovo+ results where 
not perfectly reproducible, i.e., sequencing the same mgf file twice did not 
result in identical results. Very similar of course, but not identical.

We also had a look at UniNovo (made by the same group who made PepNovo+) but it 
does not support variable modifications (and carbamidomethyl was hard coded), 
so that was not an option.

Harald

Original comment by [email protected] on 27 Mar 2014 at 4:03

[GoogleCodeExporter]

Thanks for feedback.

According to my first results, I dont think I would be using Direct Tag, as
short sequences are not very diagnostic, and irrelevant for Blasting (I'm
not running any other search programs which use (or accept) tags for
protein ID). I'm running the Fasta option now... waiting for results :-)

Alex

Original comment by [email protected] on 27 Mar 2014 at 4:26

[GoogleCodeExporter]

Harald,

You might want to check.... I just run Pepnovo+ only (had DircTag
unchecked), but results still show DireTag hits. Could you check if this is
a glitch on my system or you can reproduce this?

Original comment by [email protected] on 27 Mar 2014 at 8:42

[GoogleCodeExporter]

Issue reproduced, fixed and a new version released. :)

Harald

Original comment by [email protected] on 27 Mar 2014 at 9:53

[GoogleCodeExporter]

Harald,

Could you please expand on what is meant by Use Spectrum Precursor mz and
Use Spectrum Charge options? I read the Help menu, but it is still a little
unclear how these are used (or not) in the de-novo sequencing algorithm. I
have files for both low res MS/MS (IT scans on a VelosPro) and high
resolution HCD scans on the same instrument... just wondering which options
are best (if any difference) for both type of scans.
Also, does pepnovo correct for the charge states of the of the fragment
ions? I heard of a node for ProteinDiscoverer that can correct for fragment
charge state and deconvolute into single charge state product ion
spectra... have you tried this? do you know if that would improve the
outcomes of the denovo process??

Original comment by [email protected] on 28 Mar 2014 at 12:44

[GoogleCodeExporter]

Hi Alex,

I guess that some of these questions are better forwarded to the developers of 
PepNovo and DirecTag, but I'll do my best. The 'Use Spectrum Precursor mz' and
'Use Spectrum Charge' options basically decides if the algorithms are to use 
the values found in the mgf file for each spectrum, or if the precursor mz and 
charge should be estimated/calculated by the algorithms.

As for the difference between low resolution and high resolution instruments 
I'm not sure. I guess you just have to test. Or contact the developers of the 
algorithms. We "only" made the algorithms easily available, but are not experts 
in the algorithms themselves.

If PepNovo corrects for the charge states of the fragment ions is again a 
question for the PepNovo developers, but in our experience with database search 
engines the results are usually similar with and without deconvolution. But 
again, the easiest option would be to test both. :)

Harald

Original comment by [email protected] on 28 Mar 2014 at 1:20

[GoogleCodeExporter]

Ok, got it... basically, do you trust the mgf files precursor and charge,
or do you want pepnovo to do its own assessment.

Thanks! do you have the contact info of any of the developers?

Everything working smoothly now.

alex

Original comment by [email protected] on 28 Mar 2014 at 1:33

[GoogleCodeExporter]

Harald,

I run a .mgf file de-novo sequencing, and everything went smoothly. I now
tried the Peptide Matches (Beta) option, added a fasta file, and
modifications (there was one custom made modification). When the process
finished, I'm back to DeNovoGui's main page, but there is no report of the
Peptide Matches. In fact, the system seems to have frozen, since I can not
access any of the 'buttons' on the UI.

I can provide you with the files I have used and see whether you can
reproduce the results I've seen.

Thanks!

Alex

Original comment by [email protected] on 31 Mar 2014 at 1:04

[GoogleCodeExporter]

Harald,

Just downloaded 1.3.4. I tried Export>Blast, set a score and number of
hits, and after a select a filename to save, I get "An Error Occurred While
Exporting Results" window (and process finishes). I remember doing Blast
export in previous version of DenovoGui but dont remember getting this
alert!

The following is the bug report:

Thu Apr 03 12:01:42 ADT 2014: DeNovoGUI version 1.3.4.
java.lang.ClassCastException:
com.compomics.util.experiment.identification.PeptideAssumption cannot be
cast to com.compomics.util.experiment.identification.TagAssumption
at
com.compomics.denovogui.io.TextExporter.exportBlastPSMs(TextExporter.java:358)
at com.compomics.denovogui.gui.ResultsFrame$46.run(ResultsFrame.java:2055)
java.lang.ClassCastException:
com.compomics.util.experiment.identification.PeptideAssumption cannot be
cast to com.compomics.util.experiment.identification.TagAssumption
at
com.compomics.denovogui.io.TextExporter.exportBlastPSMs(TextExporter.java:358)
at com.compomics.denovogui.gui.ResultsFrame$46.run(ResultsFrame.java:2055)

Original comment by [email protected] on 3 Apr 2014 at 3:07

[GoogleCodeExporter]

Hi Alex,

Sorry, it seems that DenovoGUI does not like that we export data to blast after 
having used our database mapping. I will correct that right away.

Thanks for the update.

Marc

Original comment by [email protected] on 3 Apr 2014 at 3:18

[GoogleCodeExporter]

Hi Alex,

I just released DeNovoGUI v1.3.5 which should solve this issue. Please let us 
know if this is not the case.

BTW, it would be nice if you added new issues for future problems and 
suggestions. Instead of just extending this single thread, which has already 
been set to "fixed" when the first problem was solved. :)

Harald

Original comment by [email protected] on 3 Apr 2014 at 6:39