Re: [cml/ccml-discuss] Identifiers in CML

[Oliver S. <rev...@us...>]

Hi Dan, Peter,

Ah, now it gets interesting...

​Yes, we can use CML do describe a lot of different kinds of chemistry and
for each we might want/need to be able to use different kinds of
identifiers.

But I think that could be done by using the <identifier> tag with
appropriate ​conventions and/or dictionaries.

​​​I'm trying to represent structures from QM calculations, therefore I
need structural identifiers.
And for other applications we could define other conventions  or
dictionaries.

How about this:

*<cml*  xmlns="http://www.xml-cml.org/schema"
convention="convention:molecular"
xmlns:convention="http://www.xml-cml.org/convention/"
xmlns:structID="http://www.xml-cml.org/dictionary/structualIdentifier/"
     xmlns:compndID="http://www.xml-cml.org/dictionary/compoundIdentifier/"
     xmlns:substanceID="ttp://www.xml-cml.org/dictionary/substanceIdentifier/"
*>*
  *<molecule* id="aspirin" spinMultiplicity="1" formalCharge="0"*>*
    *<identifier*
dictRef="structID:InChI"*>*InChI=1S/C9H8O4/c1-6(10)13-8-5-3-2-4-7(8)9(11)12/h2-5H,1H3,(H,11,12)*</identifier**>*
    *<identifier*
dictref="structID:CanonicalSmiles"*>*CC(=O)OC1=CC=CC=C1C(=O)O*</identifier**>*
    *<identifier* dictref="compndID:pubChemCompound"*>*CID 2244*</identifier**>*
    *<identifier* dictref="​substsID:pubChemSubstance"*>*SID
53788943*</identifier**>*
    <!-- atomArray, bondArray, etc. -->
    *<property* dictRef="cml:molmass"*>*
      *<scalar* dataType="xsd:double" units="unit:dalton"
xmlns:unit="http://www.xml-cml.org/unit/si/"*>*180.15742*</scalar**>*
    *</property**>*
    *<formula* concise="C 9 H 8 O 4"*/>*
  *</molecule**>**</cml**>*

​
The dictionaries with a more general applicability could be part of the CML
standard and hosted at the CML website
and if necessary other very specific dictionaries could even be ​located
somewhere else (e.g. at NIH).


What do you think?


Best,
Oliver



On Mon, Nov 17, 2014 at 5:32 PM, Peter Murray-Rust <pm...@ca...> wrote:

> Absolutely right Dan,
>
> We never confuse
> ​​
> substance with compound with structure.
>
> P.
>
>
> On Mon, Nov 17, 2014 at 3:46 PM, Zaharevitz, Daniel (NIH/NCI) [E] <
> Dan...@ni...> wrote:
>
>> It has been quite a while, but I did use the <identifier> tag for a lot
>> of stuff. I would be interested in participating in reviving it. One reason
>> I used it was the tag name implied  a distinction from structure. I think
>> it is very important to keep the distinction between substance or sample
>> identifier distinct from structure. Chemical structure is an empirical
>> property of a substance hence subject to change as more data become
>> available. Making structure serve as a substance identifier is a major
>> mistake and thus I would argue that all the uses in the example listed are
>> inappropriate uses of the identifier tag. Of course as Peter says any
>> internally consistent use should be able to be changed to any other
>> internally consistent use without too much difficulty, but I think using
>> structure as an identifier can lead to major problems in maintaining
>> internal consistency.
>>
>> DanZ
>>
>> --
>>
>> /**********************/
>>     Daniel Zaharevitz
>>     Chief, ITB, DTP, DCTD
>>     National Cancer Institute
>>     Zah...@ma...
>> /**********************/
>>
>> From: Peter Murray-Rust <pm...@ca...<mailto:pm...@ca...>>
>> Date: Monday, November 17, 2014 3:23 PM
>> To: Oliver Stueker <rev...@us...<mailto:
>> rev...@us...>>
>> Cc: "cml...@li...<mailto:
>> cml...@li...>" <cml...@li...
>> <mailto:cml...@li...>>
>> Subject: Re: [cml/ccml-discuss] Identifiers in CML
>>
>> Wonderful.
>>
>> Egon wants to hand on the baton of CML so I'll put you in touch with
>> others quite shortly. I'd say do whatever seems reasonable and not too
>> complex...
>>
>>
>> Happy to see extensions in principle - suggest what you want to do...
>>
>> On Mon, Nov 17, 2014 at 3:12 PM, Oliver Stueker <
>> rev...@us...<mailto:rev...@us...>>
>> wrote:
>> Thanks Peter,
>>
>> My goal is to add SMILES and InChI identifiers to the molecule elements
>> in CML/CompChem documents.
>> The CompChem convention states that <molecule> elements should conform to
>> the molecular convention.
>> In the  Molecular convention there is a list of allowed child elements,
>> of which <property>, <label> and a bit far-fetched <name> come the closest
>> to describe an identifier.
>>
>> I can easily just invent something but I prefer to follow an already
>> defined standard or involve with defining and accepted standard.
>>
>> In the cmllite-validator-code repo I found this CML/CompChem file [1]
>> which defines:
>>
>> ​​
>> <cml>
>>
>>   <module role="joblist">    <identifier
>> convention="chemid:EmpiricalFormula" value="CCl2O2"/>    <identifier
>> convention="chemid:InChI" value="InChI=1/CCl2O2/c2-1(4)5-3"/>
>> <identifier convention="chemid:CanonicalSmiles" value="ClOC(=O)Cl"/>
>> <identifier convention="chemid:IsomericSmiles" value="C(=O)(Cl)OCl"/>
>> <module role="job" title="job1">
>>
>>       <!-- ... -->
>>
>>     </module>
>>
>>   </module>
>>
>> </cml>
>>
>> However on http://www.xml-cml.org/convention/ there is no chemid
>> convention nor an <identifier> element in any convention.
>>
>> What do you think of adding <identifier> to the allowed elements of
>> <molecule> in the molecular convention and starting a chemid dictionary?
>>
>>
>> In fact I'm currently also working on expanding the CompChem dictionary.
>> I've (privately) forked the repo on bitbucket.org/cml/dictionary-compchem
>> <http://bitbucket.org/cml/dictionary-compchem> and will propose a pull
>> request at a suitable time.
>> I could do the same for the molecular convention.
>>
>> Best,
>>  Oliver
>>
>> [1]
>> https://bitbucket.org/cml/cmllite-validator-code/src/2eaa18f959bb0324268bf75be8a904d4c9e07944/src/test/resources/org/xmlcml/www/convention/cmlcomp/valid/two-jobs.cml?at=default
>>
>> On Mon, Nov 17, 2014 at 3:32 PM, Peter Murray-Rust <pm...@ca...
>> <mailto:pm...@ca...>> wrote:
>> Copying Egon,
>>
>> We did have an <identifier> label which allowed this, but you can also
>> use <label> and add a "class" attribute.
>>
>> The key approach of CML now is that communities should create conventions
>> that work for them. As these become established then conventions can become
>> normalised. trying to constrains too rigidly requires a lot of software are
>> consistent discipline.
>>
>> If different communities end up with slightly different approaches it's
>> not normally hard to convert or merge.
>>
>> P.
>>
>>
>> On Mon, Nov 17, 2014 at 8:43 AM, Oliver Stueker <
>> rev...@us...<mailto:rev...@us...>>
>> wrote:
>> Dear CML Community,
>>
>>
>> ​is there an (official) standard or best practice on how to include one
>> or more identifiers (SMILES, InChI, etc.) in a CML document following the
>> molecular convention?
>>
>> Maybe as <label> or <property> children of the <molecule> ?
>>
>>
>> I couldn't find anything in the convention or the cml dictionaries.
>>
>>
>> Best,
>> Oliver ​
>>
>>
>> ​​
>> ​
>> --
>> Oliver Stueker, Dr. rer. nat.
>> Postdoctoral Fellow, Poirier Lab
>> Department of Chemistry, Memorial University
>> Room C3052  -  phone: +1 (709) 864-8752<tel:%2B1%20%28709%29%20864-8752>
>>
>>
>> ------------------------------------------------------------------------------
>> Download BIRT iHub F-Type - The Free Enterprise-Grade BIRT Server
>> from Actuate! Instantly Supercharge Your Business Reports and Dashboards
>> with Interactivity, Sharing, Native Excel Exports, App Integration & more
>> Get technology previously reserved for billion-dollar corporations, FREE
>>
>> http://pubads.g.doubleclick.net/gampad/clk?id=157005751&iu=/4140/ostg.clktrk
>> _______________________________________________
>> cml-discuss mailing list
>> cml...@li...<mailto:
>> cml...@li...>
>> https://lists.sourceforge.net/lists/listinfo/cml-discuss
>>
>>
>>
>>
>> --
>> Peter Murray-Rust
>> Reader in Molecular Informatics
>> Unilever Centre, Dep. Of Chemistry
>> University of Cambridge
>> CB2 1EW, UK
>> +44-1223-763069<tel:%2B44-1223-763069>
>>
>>
>>
>>
>> --
>> Peter Murray-Rust
>> Reader in Molecular Informatics
>> Unilever Centre, Dep. Of Chemistry
>> University of Cambridge
>> CB2 1EW, UK
>> +44-1223-763069
>>
>
>
>
> --
> Peter Murray-Rust
> Reader in Molecular Informatics
> Unilever Centre, Dep. Of Chemistry
> University of Cambridge
> CB2 1EW, UK
> +44-1223-763069
>