Re: [Psidev-ms-dev] PSI-MS controlled vocabularies proposal

[Eric D. <ede...@sy...>]

Hi David, Matt, thanks for this summary and discussion. I look forward to
the discussion in Heidelberg. I think this is all good, although I might
suggest doing updates in smaller chunks rather than one grand update, but we
can debate this.

Thanks!
Eric


> -----Original Message-----
> From: Matthew Chambers [mailto:mat...@gm...]
> Sent: Tuesday, April 05, 2011 9:49 AM
> To: psi...@li...
> Subject: Re: [Psidev-ms-dev] PSI-MS controlled vocabularies proposal
> 
> Xcalibur is perhaps misclassified, but aren't the classifications
> themselves sound? In other words,
> the structure itself is fine, but the way some of the terms are assigned
> into it may not be. I think
> there are plenty of misclassifications to be cleaned up, but the hierarchy
> seems more or less ok.
> 
> The way we get new terms into the CV is more inconsistent and represents
> the bigger problem IMO. For
> "editorial dispatch" I mean the way that journals send out papers to
> subfield experts because the
> journal doesn't staff an expert editor for every subfield. Perhaps we can
> just start sending
> unsolicited review requests to various experts. ;)
> 
> -Matt
> 
> 
> On 4/5/2011 11:14 AM, David Ovelleiro wrote:
> > Dear Matt,
> >
> > I'll be glad to send you our example material. I'll assume that the
> extension of your
> > email(m*a*t*t.c*h*a*m*b*e*r*s*42) is gmail.
> >
> > I'll send you (tomorrow, maybe in two days) a couple of attachments with
> the information. We'd
> > really appreciate knowing your opinion about this.
> >
> > I agree with you when you say that sometimes, you can find things out of
> the scope of the PSI-MS cv.
> > that is something that we'll need to handle.
> >
> > On the other hand, I don't agree when you say that there's little to
> control in software. Maybe is
> > one of the simplest sets of information, but if we are going to use the
> c.v. to validate data, we
> > need some criteria. for example: in software, you'll find the Xcalibur
> software in "adquisition",
> > "processing" and "analysis". everywhere. And, although Xcalibur is
> clearly an "adquisition"
> > software, performing several "processing" steps. I wouldn't say that
> XCalibur is also an "analysis
> > software". If you allow this, further steps of validation are impossible
> (like "what kind of
> > software are you using to adquire-process-analyse your data?"). That's
> the reason why we'll suggest
> > a clear effort on documenting things. Having this documentation, we
> would be able to know why
> > there's an LTQ-velos and not an Orbitrap-velos, for example ;)
> >
> > I'm not very sure what "editorial dispatch mechanism" means when talking
> about new terms in
> > controlled vocabularies. but if you can detail it further, we'll take it
> into account. Anyway, I
> > think that our proposal will take a simpler (and maybe more practical)
> approach for the new candidates.
> >
> > Anyway, I think it's a pity that we are not going to be able to discuss
> all this at Heidelberg. I'll
> > send you the information ASAP.
> >
> > Regards,
> >
> > David
> >
> >
> >
> >
> > On 05/04/2011 16:05, Matthew Chambers wrote:
> >> Hi David,
> >>
> >> My understanding of the main problem with adding terms to the CV is a
> lack of consensus and/or
> >> expertise. AFAIK, the largest outstanding CV request is from 7/19/2010
> from Salvador Martinez at
> >> ProteoRed. It was a big batch of terms spanning multiple subfields of
> expertise. The bulk of the
> >> terms are MALDI sample prep related and we have a decided lack of MALDI
> expertise in our working
> >> groups (AFAIK). My reply regarding those terms was:
> >>> I think most of the sample preparation terms should go
> intohttp://code.google.com/p/gelml/source/browse/trunk/CV/sep.obo.
> >>   >  Incidentally, has there been discussion about merging sep.obo with
> psi-ms.obo?
> >> This met with no comment.
> >>
> >> Some of the other terms were raw data signal processing related and I
> was the only one to comment on
> >> them which was not enough consensus to make me comfortable adding them.
> >>
> >> The software terms were quickly added by Martin because there's little
> to "control" about them.
> >>
> >> Since I'm not going to be at Heidelberg, I'd appreciate a preview of
> your detailed example changes
> >> to the software hierarchy. But I think that the CV structure is a small
> part of our CV worries. :)
> >> Perhaps we would benefit from an editorial dispatch mechanism like the
> journals use?
> >>
> >> Thanks,
> >> -Matt
> >>
> >>
> >> On 4/5/2011 7:47 AM, David Ovelleiro wrote:
> >>> Dear all,
> >>>
> >>> As announced in one of the last phone conferences, we proposed to
> >>> restructure the PSI-MS controlled vocabularies/ontology, to facilitate
> >>> maintenance as well as use in tools like the semantic validator.
> >>>
> >>> Our proposal has to main points:
> >>>
> >>> 1. Restructuration of the controlled vocabularies
> >>>
> >>> Our intention at this level can be summarized in the following points:
> >>> . Keep things as simple as possible
> >>> . Respect present id's as much as possible: we'll try to keep the
> >>> present id's in order to maintain the maximal backwards compatibility.
> >>> On the other hand, the overall structure will change. In practical
> >>> terms: that the software generating mzIdentML will work perfectly, but
> >>> the semantic validators will need to be updated.
> >>> . We'll document things: every major definition will be carefully
> built.
> >>> All definitions, when possible, will be generated using all the
> >>> information inside MIAPE specifications. References to the mzML and
> >>> mzIdentML formats will be made when possible. The proper documentation
> >>> of everything will make possible the maintenance of the controlled
> >>> vocabulary and will make things easier when semantic validators are
> built.
> >>>
> >>>
> >>> 2. Maintain and update the controlled vocabularies.
> >>>
> >>> Once the first step is finished, we'll begin with the update and
> >>> introduction of new terms proposed by the community. The exact
> mechanism
> >>> how the CV will be updated needs to be discussed in detail in
> >>> Heidelberg. I think we all agree that at the moment it is completely
> >>> impractical, too lengthy and troublesome. We need to come up with a
> new
> >>> strategy to do it
> >>>
> >>> The restructuration and update of the controlled vocabularies will be
> >>> made in a single step, sometime this year: there'll be a single update
> >>> with all the changes. In the meantime, we'll make public all the
> >>> proposals in the PSI mail list. We'll discuss with the community the
> >>> changes to be made (structure and definitions).
> >>>
> >>> A detailed example of the changes we have in mind will be shown in the
> >>> PSI meeting in Heidelberg. This example will take into account only
> the
> >>> "software" section of the controlled vocabularies.
> >>> We are expecting to hear your opinions about this proposal, and go
> into
> >>> detail in Heidelberg.
> 
> --------------------------------------------------------------------------
> ----
> Xperia(TM) PLAY
> It's a major breakthrough. An authentic gaming
> smartphone on the nation's most reliable network.
> And it wants your games.
> http://p.sf.net/sfu/verizon-sfdev
> _______________________________________________
> Psidev-ms-dev mailing list
> Psi...@li...
> https://lists.sourceforge.net/lists/listinfo/psidev-ms-dev