On Thu, Apr 30, 2015 at 10:30 AM, Tim Dudgeon <tdudgeon...@gmail.com> wrote:
> On 29/04/2015 13:39, Rajarshi Guha wrote:
>
> The pharmacophore matching was described in a CDK News issue - but all the
> links I have seem to be dead. It's very simplistic in fact - pharmacophore
> query atoms and bonds are defined as subclasses of SMARTS atoms and bonds.
> The query is parsed into a pharmacophore molecule and then a match is
> identified via subgraph isomorphism. Currently only distance and angle
> constraints are supported. The code really should be extended to support
> dihedrals and excluded volumes
>
> So this isn't clear to me. "subgraph isomorphism" makes is sound like its
> using 2D pharmacophores, whereas "distance and angle constraints" (and what
> follows) makes it sound like its 3D pharmacophores.
>
So the use of a graph in this case is really as a data structure, rather
than a 2D or 3D representation.
Consider a simple case of a distance constraint between an Aromatic group
and an Amine using a distance range of 3A - 4.2A. This query an be
represented as a graph: Ar -- Am, where the edge has a property encoding
the distance range and the nodes are the pharmacophore groups.
Then given a target molecule we first identify each instance of an Aromatic
or Amine group. Say there are 2 of the former and 1 of the latter. We can
construct a graph by connecting all the groups present to each other. As an
edge list this would look like
Ar1 -- Am1
Ar2 -- Am1
Ar3 -- Am1
with each edge have a property that stores the distance between the two
nodes comprising the edge.
At this stage you have a query graph and a target graph and you can apply
subgraph isomorphism. (Obviously the isomorphism code supports custom
defined node and edge equivalence tests)
See Fig 2 in https://dl.dropboxusercontent.com/u/26998770/wrapper.pdf
>
> Regarding the details of the definition file - are you referring to the
> definition in this particular case? Or just, generally how to define a
> pharmacophore? If the former, it's probably best to get in touch with John
> Van Drie directly.
>
> I was wanting to know what the contents of the definition file were e.g.
> how to create a new definition.
>
I don't have a specific write up on that, but see
https://github.com/rajarshi/cdkpcore for an example definition file
--
Rajarshi Guha | http://blog.rguha.net
NIH Center for Advancing Translational Science
------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
Cdk-user mailing list
Cdk-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/cdk-user
