Wrong atom types assignements

View: New views
2 Messages — Rating Filter:   Alert me  

Parent Message unknown Wrong atom types assignements

by Aziz Yasri-2 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Some parts of this message have been removed. Learn more about Nabble's security policy.
Dear All,

I am using cdk-1.2.3 version and trying to get forcefield parameters for some molecules. When I use the forcefieldConfigurator calss  with mmff94 or mm2 forcefield, I get wrong atom type names assignements. I have done this several times for many molecules and I get to the same conclusion. I have atached here  the code (for one molcule) I am using that configures the forcefield, assigns the atom type names and print the atom type names for each atom in the molecule. I also attached the atom types obtained from the MMFF94 forcefield (REAL MMFF94) from other source and I know that there are correct.

I don't know if I am doing someting wrong or the routine that assigns the atome types is not correct.


here is the code:

    sp = SmilesParser(DefaultChemObjectBuilder.getInstance())
    mol = sp.parseSmiles("CN1CC2=C(N=CN2C2=C(C1=O)C(Cl)=CC=C2)C1=NOC(=N1)[C@](C)(O)CO")
    ffc =  ForceFieldConfigurator();
    ffc.setForceFieldConfigurator("mmff94");
    ffc.assignAtomTyps(mol);
    atoms = mol.atoms().iterator()
    atom_ID = 1
    while atoms.hasNext():
        atom_i = atoms.next()
        print atom_ID,"Symbol :",atom_i.getSymbol()," Atom Type :",atom_i.getAtomTypeName()
        atom_ID += 1



Mol code : CID10091748
Mol smile : CN1CC2=C(N=CN2C2=C(C1=O)C(Cl)=CC=C2)C1=NOC(=N1)[C@](C)(O)CO


MMFF94                                REAL MMFF94
1 Symbol : C  Atom Type : C            C
2 Symbol : N  Atom Type : N            NC=O
3 Symbol : C  Atom Type : C            C
4 Symbol : C  Atom Type : Car        C5A
5 Symbol : C  Atom Type : C5        C5B
6 Symbol : N  Atom Type : N2OX        N5B
7 Symbol : C  Atom Type : C5        C5A
8 Symbol : N  Atom Type : N2OX        NPYL
9 Symbol : C  Atom Type : Car        Car
10 Symbol : C  Atom Type : Car        Car
11 Symbol : C  Atom Type : C=        C=
12 Symbol : O  Atom Type : O=        O=
13 Symbol : C  Atom Type : Car        Car
14 Symbol : Cl  Atom Type : CL        CL
15 Symbol : C  Atom Type : Car        Car
16 Symbol : C  Atom Type : Car        Car
17 Symbol : C  Atom Type : Car        Car
18 Symbol : C  Atom Type : C5        C5B   
19 Symbol : N  Atom Type : N2OX        N5A
20 Symbol : O  Atom Type : Oar        Oar
21 Symbol : C  Atom Type : C5        C5A
22 Symbol : N  Atom Type : N2OX        N5B
23 Symbol : C  Atom Type : C        C
24 Symbol : C  Atom Type : C        C
25 Symbol : O  Atom Type : O        O
26 Symbol : C  Atom Type : C        C
27 Symbol : O  Atom Type : O        O

MM2
1 Symbol : C  Atom Type : C
2 Symbol : N  Atom Type : N
3 Symbol : C  Atom Type : C
4 Symbol : C  Atom Type : Csp2
5 Symbol : C  Atom Type : Csp2
6 Symbol : N  Atom Type : NPYL
7 Symbol : C  Atom Type : Csp2
8 Symbol : N  Atom Type : N
9 Symbol : C  Atom Type : Csp2
10 Symbol : C  Atom Type : Csp2
11 Symbol : C  Atom Type : Csp2
12 Symbol : O  Atom Type : O=
13 Symbol : C  Atom Type : Csp2
14 Symbol : Cl  Atom Type : CL
15 Symbol : C  Atom Type : Csp2
16 Symbol : C  Atom Type : Csp2
17 Symbol : C  Atom Type : Csp2
18 Symbol : C  Atom Type : Csp2
19 Symbol : N  Atom Type : NPYL
20 Symbol : O  Atom Type : Oar
21 Symbol : C  Atom Type : Csp2
22 Symbol : N  Atom Type : NPYL
23 Symbol : C  Atom Type : C
24 Symbol : C  Atom Type : C
25 Symbol : O  Atom Type : O
26 Symbol : C  Atom Type : C
27 Symbol : O  Atom Type : O



Nouveau! Découvrez le Windows phone Samsung Omnia II disponible chez SFR. Cliquez ici
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Cdk-user mailing list
Cdk-user@...
https://lists.sourceforge.net/lists/listinfo/cdk-user

Re: Wrong atom types assignements

by Egon Willighagen-5 :: Rate this Message:

Reply to Author | View Threaded | Show Only this Message

Hi Aziz,

On Fri, Nov 6, 2009 at 11:17 AM, Aziz Yasri <ayasri@...> wrote:

> I am using cdk-1.2.3 version and trying to get forcefield parameters for
> some molecules. When I use the forcefieldConfigurator calss  with mmff94 or
> mm2 forcefield, I get wrong atom type names assignements. I have done this
> several times for many molecules and I get to the same conclusion. I have
> atached here  the code (for one molcule) I am using that configures the
> forcefield, assigns the atom type names and print the atom type names for
> each atom in the molecule. I also attached the atom types obtained from the
> MMFF94 forcefield (REAL MMFF94) from other source and I know that there are
> correct.
> I don't know if I am doing someting wrong or the routine that assigns the
> atome types is not correct.

I do not have experience with this part of the library, but can point
you to the unit test classes, which might server as some comparision
material. Please look at MM2AtomTypeMatcherTest [0] and
MMFF94AtomTypeMatcherTest [1]. The main difference seems to be that
these unit tests do no use the ForceFieldConfigurator, but
AtomTypeMatchers for the two force fields.

Could you try what those classes give for atom types? I think a unit
tests using those atom type matcher classes at your below mentioned
test case would be very interesting...

Egon

0.http://cdk.git.sourceforge.net/git/gitweb.cgi?p=cdk/cdk;a=blob;f=src/test/org/openscience/cdk/atomtype/MM2AtomTypeMatcherTest.java;h=647c8e6ccff5167566e3a65b572ebac7bed2de0a;hb=681b3a129f34bbbb85264bfe92451918c395960e
1.http://cdk.git.sourceforge.net/git/gitweb.cgi?p=cdk/cdk;a=blob;f=src/test/org/openscience/cdk/atomtype/MMFF94AtomTypeMatcherTest.java;h=a3f6a2a7fdca8bb01c91328889f39c35b8042b75;hb=681b3a129f34bbbb85264bfe92451918c395960e

--
Post-doc @ Uppsala University
Homepage: http://egonw.github.com/
Blog: http://chem-bla-ics.blogspot.com/
PubList: http://www.citeulike.org/user/egonw/tag/papers

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Cdk-user mailing list
Cdk-user@...
https://lists.sourceforge.net/lists/listinfo/cdk-user