if we see at the exact purpose of this discussion ,here it is...
Ok, thanks. But the getOverlaps method returns *all* MCSS which can be
quite many. In my special case I am only interested in one and not even
in the structure but only in its size (molecule clustering). Is there a
way to get this information faster than determining all MCSS?
MCSS exactly means this
You have to get the maximum with respect to some other things right?
Thorsten wants to cluster the molecule based on the substructure.
So he only want to calculate that if the given structure is having a query
substructure
or some substructure within the constraint.He is not interested in MCSS.
(this is my understanding about his problem)
If I understand you correctly, our approach would lead to combinatorial
explosion, since you would first try and match the whole SMILES string, and
then proceed by removing atoms/bonds - but you need to consider various
combinations of them to fully explore the space of all possible
substructures.
We will not remove the atom and bond,with given smile we will calculate how
many matches
we are getting like
if it is 100%(means whole substructure is matched)
else if the matches are within the limit (say <90 % but >70%)
etc.
I think the current way the UIT does it is efficient (generating the
Rgraph etc)
Yes for our purpose it is doing efficient. What i want to ask is if any body
want's to get a
subgraph for a query which is not exact match with the target but has 70%
match or something (i am using this percentage just to make the discussion
easy this 70% will mean
if query has 6 bond (4 bond are matching and 2 are not) like wise.
If this is possible his problem is solved. what do you think?
On 5/4/07, Rajarshi Guha <[EMAIL PROTECTED]> wrote:
On 5/4/07, sushil ronghe <[EMAIL PROTECTED]> wrote:
>
>
>
> What do you think about making a substructure search that will give us
> only
> Maximum substructure( in faster way and not calculating the other
> substructure)
> through UIT.
>
I'm not sure I see how one can determine the MCSS without determining all
of them? You have to get the maximum with respect to some other things
right?
If I understand you correctly, our approach would lead to combinatorial
explosion, since you would first try and match the whole SMILES string, and
then proceed by removing atoms/bonds - but you need to consider various
combinations of them to fully explore the space of all possible
substructures.
I think the current way the UIT does it is efficient (generating the
Rgraph etc)
--
Rajarshi Guha
--
********************************
sushil ronghe
Center of Pharmacoinformatics
India
*********************************
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Cdk-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/cdk-user
