Re: [Open Babel] Combine two SMARTS patterns

[Andrew D. <da...@da...>]

On Mar 5, 2009, at 4:43 PM, Andreas Maunz wrote:
> I was wondering if there is a possibility to combine two distinct  
> smarts
> patterns.

Given "NCO" and "NO", what should the merged query be? Four
possibilities are:

   N1CO1   -- two atoms in common between the input SMARTS
   ONCO    -- one atom in common
   NCON    -- one atom in common
   NCO.NO  -- no atoms in common



Given  c1ccccc1O and NccC, what should the merged query be?
Possibilities where the most number of atoms are shared are:

    c1(N)cc(C)ccc1O
    c1c(N)cc(C)cc1O
    c1cc(N)cc(C)c1O
    c1ccc(N)cc1(C)O  (which is impossible chemistry)
    c1ccc(C)cc1(N)O  (which is impossible chemistry)

and with only one atom shared, the following are possible:
    c1(cC)c(N)cccc1O
    c1c(cC)c(N)ccc1O
    c1cc(cC)c(N)cc1O
    c1ccc(cC)c(N)c1O
(plus a few more)

It's possible to enumerate all maximum common substructure
between the two SMARTS, but the set of results might
contain query graphs which have no chemical meaning. I
can't see how such a set would be useful.

				Andrew
				da...@da...