All,

Since this is important topic, I made the wiki page externally visible.

http://opentox.org/dev/apis/api-1.2/substructure-api-wishlist


Would be good to have comments from everybody, working with:

1) Fragment features of any kind
Do we need specific extension of the API for fragmens , and how does it
integrates with existing modeling algorithms?  There are at least two
partners, which already use the current API for representation of fragments
and fragment based algorithms.  What / is a new extension necessary?

2) Interested in representing properties, calculated not to from a single
compound, but to two (or more)
How to extend the API , to be able to represent entities, related to two
compounds - e.g. common substructure, alignment matrix ,etc. ?

Best regards,
Nina

On 21 October 2010 17:15, Nina Jeliazkova <jeliazkova.nina at gmail.com> wrote:

> Hi Surajit, All,
>
> Let's start a wiki page on this topic and compile a "wish list" as a first
> step.
>
>
> Best regards,
> Nina
>
> On 21 October 2010 17:06, surajit ray <mr.surajit.ray at gmail.com> wrote:
>
> > Hi Nina, All,
> >
> > I was wondering if an ontological class called "substructure" can be made
> > available in the next API. The basic functionality could just be a
> replica
> > of the "compound" class and its corresponding functionality.  In addition
> > we
> > could include some other properties which would be things like
> > "ot:isPartOf"
> > etc. (we could have a discussion in the SWDT about needed properties)
> >
> > Also the atomic features can be considered to be features of a single
> atom
> > substructure. This way we could make space for atom features as well.
> >
> > Substructures have become very important in the latest in-silico model
> > making strategies. Some papers that Om discussed from the EURO Qsar
> > outlined
> > the importance of substructures in clustering groups of molecules. Our
> > ability to have substructures as an available unit of operation  (through
> > the API) - will greatly enhance the ease of integration with the recent
> > trends in in-silico small molecule testing. Right now the only way is to
> > internalize this information within models/algorithms.
> >
> > Another big usefulness  of this would be a ready audience with drug-dev
> > companies. They usually are looking for substructures in their repos (of
> > existing compounds). More often the whole upstream workflow for the drug
> > development chemistry is based on concepts of substructure recombinations
> > using tightly controlled reactions. However I must admit this benefit -
> > does
> >  have a weak correlation with our toxicity related goals.
> >
> > Overall I feel it will be a useful and very needed addition to the API.
> >
> > Thanks
> > Surajit
> >
> >
> >
> _______________________________________________
> Partners mailing list
> Partners at opentox.org
> http://www.opentox.org/mailman/listinfo/partners
>