Conformers generation

Hi.


I would like to generate conformers from a 2D structure using Clean3D.calcConformers().


I am not convinced by the ones currently obtained, my collegues stating the conformers are quite the same.


what is the best (considering I would like to have the more diversity obtainable in the Molecule array) parameter to use for Molecule.clean(3, "XXXX") ?


Yes, i have read the clean3D options of molConvert. I have the 3.1.7.1 version of Jchem. Current parameters are "S0[ca][prehydrogenise][fm]{100}{100}{100}L11121[timelimit]{2400}"


Thanks in advance for your tips.





Regards,


Fabien

Hi Fabien,





Well, diversity limit is available from Marvin 4.1.2 only through molconvert, however, I advise waitnig for the upcoming release of JChem, since it will support diversity also in the GUI and through cxcalc (calculatir plugin).





Hope this will suit to your needs.





Best wishes,





Ödön

FTillier wrote:

Hi. I would like to generate conformers from a 2D structure using Clean3D.calcConformers(). I am not convinced by the ones currently obtained, my collegues stating the conformers are quite the same. what is the best (considering I would like to have the more diversity obtainable in the Molecule array) parameter to use for Molecule.clean(3, "XXXX") ? Yes, i have read the clean3D options of molConvert. I have the 3.1.7.1 version of Jchem. Current parameters are "S0[ca][prehydrogenise][fm]{100}{100}{100}L11121[timelimit]{2400}" Thanks in advance for your tips. Regards, Fabien

Thanks for your answer Ödön.


When will the new version be available ?


Also, what do you mean by diversity limit ?


Thanks in advance





Fabien

Quote:

When will the new version be available ?

Sooner.

Hi Fabien,





Diversity limit means the following: we superimpose two conformers and calculate the RMS distance of the corresponding atoms. If it is not greater then the diversitz limit the two conformers are treated as identical and the higher energz one is removed from the list.





I hope this was clear.





Ödön

FTillier wrote:

Thanks for your answer Ödön. When will the new version be available ? Also, what do you mean by diversity limit ? Thanks in advance Fabien

Hi Ödön.


It is clearer now, thanks.


What I still don't get, is if this diversity limit thing will be available in the API, as we are using the clean and calcConformers methods in our code.





Thanks for your patience.


Fabien

Hi Fabien,





Of course, it is going to be available in the API, both in Molecule.clean or through the culculator plugin. Also, if you check "molconvert -H3D" and cannot find:





...


[diversity] Use custom diversity criterion


Argument 1: Diversity limit for conformer equivalence check


The diversity limit will be used as minimal RMSD between


valid conformers. The default value is 0.1 (Angstrom)


...





in the printout then simply download the latest Marvin release and use it's API (available from Marvin 4.1.2).





Best wishes,





Ödön

FTillier wrote:

Hi Ödön. It is clearer now, thanks. What I still don't get, is if this diversity limit thing will be available in the API, as we are using the clean and calcConformers methods in our code. Thanks for your patience. Fabien

Hi Ödön


Thanks for input.





However, I have tried several time since past week to download last version of Jchem (jchem_3_2.tar.gz) without success. Each time the archive is corrupted. I might not post in the good topic, though...

I could download jchem_3_2.tar.gz and I have managed to extract it:

Code:

tar xzvf jchem_3_2.tar.gz

If you have got a Windows machine, download jchem_3_2.zip. The tar.gz is the Linux / OS X version.

Thanks for input. Seems the problem comes from our end. Sorry for the erroneous alert.


I have a last question regarding conformer generation.


I am using the 3.2 version, and I am unable to get the calculated energy for the generated conformers. We used to get it with


Molecule.getPropertyObject("conformerinfoenergy")


I have see in the SDF that the field is "Energy", but it's value is always null, even with this string


"[prehydrogenize][ca]{100}{20}L1[E][diversity]{0.1}[timelimit]{3600}"


as options for clean.


Can you help me on that point (and after that, I stop posting on this thread :o)





Regards,


Fabien

Hi,

FTillier wrote:

I am using the 3.2 version, and I am unable to get the calculated energy for the generated conformers. We used to get it with Molecule.getPropertyObject("conformerinfoenergy") I have see in the SDF that the field is "Energy", but it's value is always null, even with this string "[prehydrogenize][ca]{100}{20}L1[E][diversity]{0.1}[timelimit]{3600}" as options for clean.

Your option string should work. Replace the

Code:

Molecule.getPropertyObject("conformerinfoenergy")

call with

Code:

Molecule.getProperty("Energy")

Regards,


Zsolt

Hi Zsolt.


It doesn't work with the Version 3.2, it seems.


I can generate conformers, but the dreiding energy is not available (Ödon says diversity limit is only available in molConvert, so I assume it may be why is it so).


And if I use Enery rather than conformerinfoenergy I have an Exception, which may be because the type of the property have changed (from double to string, just a guess).


I am a bit stick now, as I need the dreiding energy value...





Thanks in advance for your answers.


Fabien

Hi Fabien,





Please send me the non-working part of your code.





You can also use the getDreidingEnergy() method of the GeometryPlugin.





ConformerPlugin will also include a getDreidingEnergy method in the next version of Marvin.





Best regards,


Zsolt

Szolt,


I have changed my code since last post to use the ConformerPlugin, to be able to use the new version as soon as it is available with the diversity limit enabled.





Molecule[] confs = null;


ConformerPlugin CPlugin = new ConformerPlugin();


try {


CPlugin.setMolecule(d2);


CPlugin.setMaxNumberOfConformers(100);


CPlugin.setPrehydrogenize(true);


CPlugin.setOptimization(ConformerPlugin.OPTIMIZATION_NORMAL);


// we want some conformers, not one


CPlugin.setLowestEnergyConformerCalculation(false);


CPlugin.setTimelimit(3600);


//CPlugin.setDiversity(0.1);


CPlugin.run();


confs = CPlugin.getConformers();


} catch (PluginException e) {


e.printStackTrace();


}





I then try to display the properties


for(int ci = 0;ci<confs.getPropertyCount();ci++) {


System.out.print(""+ci+":"+confs.getPropertyKey(ci));


}





The only avaliable is


EnergyValue


and the value is null.


What am I doing wrong ?

Hi Fabien,





I have attached a code example (based on your code), that will work.





Usage:

Code:

java ConformerPluginTest test.sdf

Output:

Code:

$ java ConformerPluginTest test.sdf Energy of conformer 0: 21.05238243265725 Energy of conformer 1: 21.075503861537292 Energy of conformer 2: 21.27064295412772 Energy of conformer 3: 21.457567681601613 Energy of conformer 4: 25.091447956357097 Energy of conformer 5: 25.246856785978224 Energy of conformer 6: 25.314001796582104 Energy of conformer 7: 25.340691533788384 Energy of conformer 8: 27.63894478264009 Energy of conformer 9: 28.331931128990213

In the newest version of Marvin (4.1.3) and JChem (3.2.1 <- EDIT!) ConformerPlugin.setDiversity(double) method is already available.





In the next version of Marvin a

Code:

public double getEnergy(int)

will be also availabe in the ConformerPlugin API to get the energy of the i-th conformer.





Best regards,


Zsolt

I will try the code, but the version you describe is not available, to me at least...


I saw there is a 3.2.1 version that have appeared (today ?).


I may give it a try.





Regards,


Fabien

Sorry, I mean JChem 3.2.1. This method is already available in JChem 3.2.1.

ok, it is working now. I can get the Dreiding energy when using the version 3.2.1.


Thanks for your support


Fabien