I agree with Martin. I would add that peptides are often identified many
times due to redundant acquisition as well.
There should be value at the protein level as well. You have to be able to
put the full protein sequence and I wouldn't want to do that more than
once. That said, it may be reasonable to have use cases where definition
of what protein accession was identified is done by a pointer to an
external resource via URI. Thus, use of the molecule table could be
optional, but I really think you need it.
Sean
"Martin Eisenacher" <mar...@ru...>
Sent by: psi...@li...
04/29/2008 08:14 AM
To
<psi...@li...>
cc
Subject
Re: [Psidev-pi-dev] molecule table thoughts
Meanwhile I think it saves us a lot, if a peptide (with modifications) is
found in
more than one spectrum. In the SpectrumSearchResultItem there is the link
between
spectrum and peptide (with mods) TOGETHER with a score. Additionally we
have to
report the peptide sequence (with mods) only once, although it may be
found with more
than one charge states.
Bye
Martin
Von: psi...@li...
[mailto:psi...@li...] Im Auftrag von Angel
Pizarro
Gesendet: Sunday, April 27, 2008 4:46 PM
An: psi...@li...
Betreff: [Psidev-pi-dev] molecule table thoughts
More and more I don't think our molecule lookup table is gaining us
anything. Specifically, last time at the EBI, Martin and I were discussing
modifications and how best and proper to represent them and we came to the
conclusion that the modification information should live directly next to
the scoring information. Also taking into account that a protein
determination step is an analysis, we can see that representing the
protein groups can also be defined within the result set containing the
scoring information.
So the only thing that the molecule table buys us is an easy look-up for
information without context and possibly (although not proven) file size
savings. The cost is that the format MUST reference the lookup table,
adding complexity to encoding and reading the actual results.
I don't think the convenience is worth the cost. Can we move the
identified compound information to be collocated with the results
information?
-angel
-------------------------------------------------------------------------
This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
Don't miss this year's exciting event. There's still time to save $100.
Use priority code J8TL2D2.
http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
_______________________________________________
Psidev-pi-dev mailing list
Psi...@li...
https://lists.sourceforge.net/lists/listinfo/psidev-pi-dev
|
