mcs

Dear Sirs,





i have to perform a MCS on two molecules and get the atom indices of the query that matches my following code throws a nullpointerexeption because the results are null:


MCS s2= new MCS();


s2.setMolecules(query,target);


s2.setQuery(query);


s2.setTarget(target);


s2.setMCSMode(true);





s2.search();


int results[]=s2.getResult();


is there a bug and how do i interpret the results array?


plz help me .

sorry this is the hole code:


bufferedInputStream tis1=null;


tis1 = new BufferedInputStream(


new FileInputStream("referenceMolecule.sdf"));


MolInputStream tmis1 = new MolInputStream(tis1);


MolImporter tmolimp1 = new MolImporter(tmis1);





Molecule query = tmolimp1.read();











// use Molfile molecule as target


BufferedInputStream tis=null;


tis = new BufferedInputStream(


new FileInputStream("reconstructedMolecule.sdf"));


MolInputStream tmis = new MolInputStream(tis);


MolImporter tmolimp = new MolImporter(tmis);


Molecule target = tmolimp.read();





MCS s2= new MCS();


s2.setMolecules(query,target);


s2.setQuery(query);


s2.setTarget(target);


s2.setMCSMode(true);





s2.search();


int results[]=s2.getResult();


if(results==null){


// setting the matchesReference flag for every not matching atom


for(int i=0;i<=s2.getResultSize();i++){


System.out.println(results);


}

Hi,





try a somewhat modified code, sg. like:

Code:

if ( s2.search() ) {     int results[]=s2.getResult();     .     . }

Does this help?





Regards,


Miklos

Hi mvargyas thanks for the reply,


but search dont work at all here ,it seems that the molecules are not loaded here .


because "s2.search()" returns false.

When search returns false it means that no solution was found, see the API documentation: http://www.chemaxon.com/jchem/doc/api/chemaxon/sss/search/MCS.html#search().





Does this help?





Try a simple case when the two structure are the same. Do you get an output?





So the code should look sg like this:

Code:

if ( s2.search() ) {     int results[]=s2.getResult();     .     . } else {     System.out.println("No MCS found."); // means that there's no MCS that meets all constraints }

Regards,


Miklos

hi mvargyas





thank you very much it is working fine now.

Hi,





i have the feeling that the api mcs ist less accurate than the binary mcs ,the results habe big differences.Or i am using the wrong parameters?Does anyone have the same experience.

Hi,





Miklos is abroad for the rest of the week, but let me give you a short answer.

Quote:

i have the feeling that the api mcs ist less accurate than the binary mcs

The program uses the very same API, therefore the same algorithm, so this is impossible.





That is, of course, if you are using the same version of the program and the API. You can check the version of a jchem.jar in the following way:





java -jar jchem.jar





Please see this topic for further information:


http://www.chemaxon.com/forum/ftopic127.html





Unfortunately might be a bit harder to determine which jchem version the mcs was released with.


1. Display a structure in the GUI


2. Right-click on it, select Help -> About Marvin


3. Go the the history of JChem changes to check which JChem contains this Marvin version:





http://www.chemaxon.com/jchem/changes.html





We will unify our product versions in the future, and determining the program version will also be more easier.

Quote:

Or i am using the wrong parameters?

If you are using the same version, then then it must be the way of usage indeed.


If you send us an example (input, source code, result difference) where the difference can be reproduced and you do not understand the reason, Miklos will take a look at your code next week and see if there are obvious problems.





Best regards,





Szilard

Hi


my jchem api version has 3.2.11


Table version: 41


my MarvinView is at version 4.0.1 is it the August 23, 2005: 3.1 version then?


my code is the following:


public class mcs {





/**


* @param args


*/


public static void main(String[] args) {


try {

















BufferedInputStream tis1=null;


tis1 = new BufferedInputStream(


new FileInputStream("reconstructedMolecule.sdf"));


MolInputStream tmis1 = new MolInputStream(tis1);


MolImporter tmolimp1 = new MolImporter(tmis1);


Molecule target1 = tmolimp1.read();








// use sdfile molecule as target


BufferedInputStream tis=null;


tis = new BufferedInputStream(


new FileInputStream("referenceMolecule.sdf"));


MolInputStream tmis = new MolInputStream(tis);


MolImporter tmolimp = new MolImporter(tmis);


Molecule target = tmolimp.read();








MCS s2= new MCS();


s2.setMolecules(target1,target);


//s2.setMCESMode(true);


s2.setMCSMode(true);


System.out.println(chemaxon.jchem.VersionInfo.JCHEM_VERSION);


System.out.println(chemaxon.jchem.VersionInfo.JCHEM_TABLE_VERSION);


if ( s2.search() ) {


System.out.println("sucht und sucht");


int results[]=s2.getResult();


for(int i=0;i<s2.getResultSize();i++){


System.out.println(results);


}


}





// MolAtom resultAtom[]= s2.getResultTargetAtoms();





// System.out.println(s2.getResultSize());








} catch (IOException e) {


e.printStackTrace();


System.exit(1);





}//end catch


}//end main


}//end searchTest





-------------------


the result is:


[7, 5, 0, 4, 3, 6, 9, -1, 1, 8, -1, -1, -1]


the result from mcs is:


1 -> 8


2 -> 6


3 -> 7


4 -> 1


5 -> 2


6 -> 5


7 -> 4


8 -> 3


9 -> 10


10 -> 9


11 -> 11


12 -> 12





the sdfiles are attached.


thank you very much!

Hi ,also when i try to iterate through getResultqueryAtoms


for(MolAtom ma : s2.getResultQueryAtoms()) {


logger.info(ma);


}





i get the following :


java.lang.ArrayIndexOutOfBoundsException: 9


at chemaxon.marvin.modules.CommonStructureSearch.getResultQueryAtoms(CommonStructureSearch.java:664)


at


chemaxon.sss.search.MCS.getResultQueryAtoms(MCS.java:302)

I will check both problems you reported and then get back to this forum to provide detailed and specific response asap.





regards,


Miklos

ok i am looking forward to it.


thanks a lot

Hi mvargyas





did you have any results ?


my tast kind of depend on this :-)





thanx

Hi there,





apologies for not getting back to you sooner.


Could you check with the recent release version 5.0? I believe the problem has been fixed. It certainly does not crash, and the output of your sample code is:

Quote:

sucht und sucht 7 5 6 0 1 4 3 2 9 8 10 11

However, please note the the API of the MCS class has slightly been changed that also affects your code. In particular, the return type of method getResultQueryAtoms() is int[] (not MolAtom[] any more), see: http://www.chemaxon.com/jchem/doc/api/chemaxon/sss/search/MCS.html#getResultQueryAtoms().





Does this solve your problem?





Regards,


Miklos

works like a charm thank you very much :-)

Thanks for your patience! Please report all problems you encounter, also, all suggestions and feature requests are welcome.





Miklos

Hi Miklos and Mvargyas





i just noticed that the getRestultQueryBonds() method always returns null


whether there is a result as int[][] from getResult() or not.


Its seems to be a bug.





Greets


msouiai

Hi Msouiai,





indeed, I can confirm that it is a bug. We will fix it in the next minor release (5.0.1, due around 4 February).





Regards,


Miklos

According to the most recent information I got the next bugfix release is scheduled on 13 February.


Miklos

mvargyas wrote:

According to the most recent information I got the next bugfix release is scheduled on 13 February. Miklos

Please update me


karthik

The bug has been fixed and tested.


The release is still scheduled by Wednesday, 13 February. Alpha release will be build on Monday.





Miklos

Hi Karthik,

Quote:

Please update me

Please "Watch this topic for replies" (link at top and bottom) to get a notification each time a new release comes out (either major or minor):





http://www.chemaxon.com/forum/ftopic287.html





The same topic for Marvin:





http://www.chemaxon.com/forum/ftopic290.html





For Instant JChem:





http://www.chemaxon.com/forum/ftopic1728.html





Best regards,





Szilard

Hi Szilard,


thanks for the bugfix ,i cant see the release of 5.01 in the download section though.


regards


msouiai

The patch release is not out yet (though it was scheduled for yesterday). We expect it by 1300 hour today.... You'll be notified.





Miklos

Release of jchem was delayed due to a critical bug in marvinsketch.


We will notify you when he stable version is released.





M.

JChem 5.0.1 has been released.


Regards,


Miklos

Hi Miklos,





i tried the fixed mcs module now, and there is no nullpointer exeption anymore.


But i got wrong results ,for instance with the example i posted above i got following:


edge between:6and6


edge between:7and6


edge between:6and6


edge between:6and6


edge between:6and6


edge between:6and6


edge between:6and6


edge between:6and6


edge between:6and6


edge between:6and6


edge between:6and8


edge between:6and8





and that cant be :-(

Hm... that's very odd indeed. We'll check it and get back to you soon.


Regards,


Miklos

Hi Miklos ,


thanks for your patience with me,i have another quetion ,


is there a possibility to search for the actual mcs not the mces? I want to find also not connected components.


Regards


Msouiai

Hm, this is puzzling: I tried your structures and I got this output:





query bonds:


0 - 1


5 - 0


1 - 2


2 - 3


3 - 4


4 - 5


5 - 6


5 - 7


6 - 9


7 - 8


9 - 10


9 - 11





target bonds:


5 - 6


1 - 5


6 - 7


4 - 7


3 - 4


1 - 3


2 - 1


0 - 1


2 - 8


9 - 0


8 - 10


8 - 11





This looks o.k. though I did not check each and every bond one-by-one just picked some of them...


What do you think?





Thanks


Miklos

Hi , this is my code:


import java.io.BufferedInputStream;


import java.io.FileInputStream;


import java.io.IOException;





import chemaxon.formats.MolImporter;


import chemaxon.formats.MolInputStream;


import chemaxon.sss.search.MCS;


import chemaxon.sss.search.SearchException;


import chemaxon.sss.search.StandardizedMolSearch;


import chemaxon.struc.MolAtom;


import chemaxon.struc.MolBond;


import chemaxon.struc.Molecule;


import chemaxon.util.MolHandler;








public class mcs {





/**


* @param args


*/


public static void main(String[] args) {


// TODO Auto-generated method stub


try {





BufferedInputStream tis1=null;


tis1 = new BufferedInputStream(


new FileInputStream("reconstructedMolecule.sdf"));


MolInputStream tmis1 = new MolInputStream(tis1);


MolImporter tmolimp1 = new MolImporter(tmis1);


Molecule target1 = tmolimp1.read();


target1.aromatize();





// use sdfile molecule as target


BufferedInputStream tis=null;


tis = new BufferedInputStream(


new FileInputStream("referenceMolecule.sdf"));


MolInputStream tmis = new MolInputStream(tis);


MolImporter tmolimp = new MolImporter(tmis);


Molecule target = tmolimp.read();


target.aromatize();





MCS s2= new MCS();


s2.setMolecules(target1,target);


//s2.setMCESMode(true);








if ( s2.search() ) {





MolBond[] resultBonds=s2.getResultQueryBonds();


for (int i = 0; i < resultBonds.length; i++) {


System.out.println("bond"+i+":"+resultBonds.getAtom1().getAtno()+"and"+resultBonds.getAtom2().getAtno());


}








}











} catch (IOException e) {


e.printStackTrace();


System.exit(1);





}//end catch


}//end main


}//end searchTest








my result is once again:


bond0:6and6


bond1:7and6


bond2:6and6


bond3:6and6


bond4:6and6


bond5:6and6


bond6:6and6


bond7:6and6


bond8:6and6


bond9:6and6


bond10:6and8


bond11:6and8





what did i do wrong?


greets


msouiai

Hi, I guess you did not do anything wrong. The difference you encountered between my output and yours (and possible between the output you got and the one you expected, supposedly) is due to the fact that you printed atno()-s, that is, atomic numbers, but not actual atom indeces in your molecules. You got many 6-s as you have some Carbon atoms in you structure.





To obtain indeces you need sg like this:

Code:

if ( mcs.search() ) {     System.out.println( mcs.getResultAsMolecule().toFormat( "smiles"));      System.out.println( "query: " );      MolBond [] qb = mcs.getResultQueryBonds();      for ( int i = 0; i < qb.length; i++ ) {          MolAtom a1 = qb[i].getAtom1();          MolAtom a2 = qb[i].getAtom2();          System.out.println( query.indexOf(a1) + " - " + query.indexOf(a2) );      }      System.out.println( "target: " );      MolBond [] tb = mcs.getResultTargetBonds();      for ( int i = 0; i < tb.length; i++ ) {          MolAtom a1 = tb[i].getAtom1();          MolAtom a2 = tb[i].getAtom2();          System.out.println( target.indexOf(a1) + " - " + target.indexOf(a2) );      }                         }

Does this help?





Regards,


Miklos

Hi Miklos ,


yeah your right , it was my fault ,everythings working fine now.


thanks again :-)


mohamed souiai

I'm glad it works now.


Please come back to the forum if you have any further enquiry.


Miklos

Hi Miklos;


i found something interessting the mcs.search () dont terminate at the following input attached.


regards


msouiai

I will check this for you later today. However, the search options would be useful to know (e.g. atom and bond types were matched or ignored, exact, fast or turbo mode was applied etc.)





Thanks


Miklos

Hi Miklos,





i haven't touched the options so everything should be in default mode.





regards





msouiai

Hi,





huuuhh, these structures are monsters! :-)





No wonder it takes long to compute the MCS. Try the turbo mode by adding the -m turbo option to the command line. It does find a large common core in one second, though there is no guarantee that it is the largest one. In case of your particular structures, though, I reckon (by 'eye inspection' only!) that it is the exact MCS.

Code:

miklos-vargyas-computer:~/cvs/release/5.0/jchem mico$ time ./jchemsite/bin/mcs -q ~/work/test/data/mcs/referenceMolecule2.sdf -t ~/work/test/data/mcs/reconstructedMolecule2.sdf  -m turbo OCCOCCOCCOCC(COCCOCCOCCO)Oc1cc(cc(OC(COCCOCCOCCO)COCCOCCOCCO)c1OCCCOC(=O)CCC(C(O)=O)C(O)=O)C(=O)NCCCc1cc(CCCNC(=O)c2cc(OC(COCCO)COCCO)c(OCCCNS)c(OC(COCCO)COCCO)c2)cc(c1)C(O)=O real    0m1.188s user    0m0.988s sys     0m0.101s

Does this help?





Regards,


Miklos

Hi Miklos ,


yeah the turbo mode ist working finde with that example :-)


thanks

I'm glad! ;-)


M