Regarding libmcs

Hi,


Is it possible to get the various fields of the SDF file in the output of libmcs? I know that the -w option is still under development, but can this be done using the command line option? I want to get the molecular Id since it will be easier for me to track the structures that are clustered together.





Thanks in advance,





Subbarao

Hi Subbarao,





I am afraid this is not yet supported.


I will post a note here when SDFile fields are processed and printed in the output.





Regards,


Miklos

Dear Miklos,


I am having problem with the smiles format the libmcs is giving. Since i need to track back and see which molecules are coming in one cluster, i tried babel for converting into smiles. But the smile format was not the same. Then I tried molconvert in jchem and it gave the smiles with Hydrogens. I was planning to write a script which makes a pattern match for the smiles. Please tell me how to get the smiles format as seen in libmcs output.





Regards,





Subbrao

Hi,





To get the same result in molconvert try smiles:-H which should do the trick.





Please get back to the forum if this does not solve your problem.





Regards,


Miklos

Hi,


Instead of giving the .sdf file as input to Libmcs, I tried the following :


1. Use molconvert to convert sdf to smiles and do libmcs, since hydrogens were added the number of clusters changed.


2. Use molconvert to convert sdf to smiles (with smiles:H option), the number of clusters were almost same (with sdf I got 188 and with smiles I got 187).


But when I tried finding a pattern match I couldnt get the match. What should I do?





Regards,





Subbarao

Do you mean that common substructures found by libmcs when input structures came from smiles did not match results when input was taken from sdf? I'll check this and get back to you...


Miklos

I checked if different input file types give different result or not. I found that results are exactly the same.


Please explain me what do you mean by not finding a match.


Thanks


Miklos

Dear Miklos,


Thanks for the reply. What I meant was that the smile format of the input and ourput of libmcs are not matching, hence I couldnt find a match. Like for example I ran a test smiles, converted from sdf to smiles using molconvert, which is as follows:


CC1=NOC2=C1C(SCC(=O)C3=CC=C(Cl)C=C3)=NC=N2


CCOC(=O)C1=CSC(NC(=O)CSC2=C(Cl)C=CC(Cl)=C2)=N1





The output of libmcs is as follows:


Level 0


Cc1noc2ncnc(SCC(=O)c3ccc(Cl)cc3)c12 0 1


CCOC(=O)c1csc(NC(=O)CSc2cc(Cl)ccc2Cl)n1 1 1





Since I doing exact matches in my script I am having problems with this output.





Regards,





Subbarao

nsrao_jnu wrote:

Since I doing exact matches in my script I am having problems with this output.

Do you mean string matching? I recommend to aromatize the structures during conversion from sdf:

Code:

molconvert smiles:a-H input.sdf -o output.smi

See the following link for all smiles export switches:





http://www.chemaxon.com/marvin/doc/user/smiles-doc.html#options





Alternatively, you can use the chemical search tool jcsearch to match molecules. Jcsearch takes care of aromatic rings. (Similar to libmcs)





This example matches two of your smiles strings in exact matching mode:

Code:

$ jcsearch -t:e -q 'CC1=NOC2=C1C(SCC(=O)C3=CC=C(Cl)C=C3)=NC=N2' 'Cc1noc2ncnc(SCC(=O)c3ccc(Cl)cc3)c12' Cc1noc2ncnc(SCC(=O)c3ccc(Cl)cc3)c12

Jcsearch produces output only if there is a hit.





This link points to the documentation:


http://www.jchem.com/doc/user/Jcsearch.html





All the best,


Szabolcs

Hi,


Thanks for the reply. I am sorry since I confused you with the words "exact match". Yes I want to do string matching. I am able to do it now. Thanks


:)


Regards,





Subbarao

Dear Myklos,





I've read previous posts in this thread, and i'm also interested in knowing the estimated release date of a version of libmcs (i'm using 3.0.14 now) that is able to display SDF labels in its graphical interface. It would be very useful for identifying molecules inside each cluster.





On the other side, i found that in the libmcs graphical window, menus Settings and View just don't have anything inside them, and menu File options' Open, Save and Print just don't do anything at all. I suppose this is because it's still in development, am i right?





Thank you,





Guillem Plasencia.

Hi Guillem,








yes, you are right: libMCS is under development. Serious reconstruction of the GUI will be started next week. It is a few months works. After that SDF label display as well as filtering/colouring by SDF tags (like activity) will be available in LibMCS.


Will also try to integrate R-Group decomposition into LibMCS.





Regards,


Miklos