conformer generation w/o stereo OK, with R/S fails

Hi,

same molecule, but one with RS fails in conformergeneration, both with cxcalc

and with molconvert. 1st is ok, second fails.

CC1C2CC3=C(C1(CCN2CCC4=CC=CC=C4)C)C=C(C=C3)O

[H]OC1=C([H])C2=C(C([H])=C1[H])C([H])([H])[C@]1([H])N(C([H])([H])C([H])([H])C3=C([H])C([H])=C([H])C([H])=C3[H])C([H])([H])C([H])([H])[C@]2(C([H])([H])[H])[C@@]1([H])C([H])([H])[H]

names

1,13-dimethyl-10-(2-phenylethyl)-10- azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4- ol

(1R,9S,13R)-1,13-dimethyl-10-(2-phenylethyl)-10- azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4- ol

I used molconvert -3:L3 SDF problem.smi

and

cxcalc conformers problem.smi -y -O 3 -m 1 -f sdf

 

First mol is generated in a second, the other will not finish. SMI23D created correct conformers for that mol. The H enrichde molecule came from Standardizer.

Tobias

Dear Tobias,

 

Could you attach the 3D structure generated by SMI23D?

 

regards,

Gabor

_

Hi,

2d Mols are:

1,13-dimethyl-10-(2-phenylethyl)-10- azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4- ol

(1R,9S,13R)-1,13-dimethyl-10-(2-phenylethyl)-10- azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4- ol

3D Mols optimized by SMI23D  are see (problem-3D-opt-smi23d.sdf):

(1R,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10- azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4- ol

(1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10- azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4- ol

So there is a switch in absolute configuration for the second one.

A totoal of 8 stereosiomers can be generated from CC1C2CC3=C(C1(CCN2CCC4=CC=CC=C4)C)C=C(C=C3)O

I get an error when I try to generate stereoisomers the first SMILES with Marvin 5.2.2. and 3D check on.

java.lang.NullPointerException
    at chemaxon.marvin.calculations.MultipleMoleculeDisplay.getResultComponent(MultipleMoleculeDisplay.java:254)
    at chemaxon.marvin.plugin.CalculatorPluginManager.displayResult(CalculatorPluginManager.java:223)
    at chemaxon.marvin.plugin.CalculatorPluginManager.runPlugin(CalculatorPluginManager.java:184)
    at chemaxon.marvin.plugin.CalculatorPluginManager.runPlugin(CalculatorPluginManager.java:106)
    at chemaxon.marvin.plugin.CalculatorPluginManager.callback(CalculatorPluginManager.java:76)
    at chemaxon.marvin.plugin.PluginManager.callback(PluginManager.java:87)
    at chemaxon.marvin.common.swing.MolPanel$5.run(MolPanel.java:3874)

SMILES generated with Marvin:

C[C@@H]1[C@@H]2CC3=C(C=C(O)C=C3)[C@]1(C)CCN2CCC1=CC=CC=C1
C[C@H]1[C@@H]2CC3=C(C=C(O)C=C3)[C@]1(C)CCN2CCC1=CC=CC=C1
C[C@@H]1[C@H]2CC3=C(C=C(O)C=C3)[C@]1(C)CCN2CCC1=CC=CC=C1
C[C@H]1[C@H]2CC3=C(C=C(O)C=C3)[C@]1(C)CCN2CCC1=CC=CC=C1
C[C@@H]1[C@@H]2CC3=C(C=C(O)C=C3)[C@@]1(C)CCN2CCC1=CC=CC=C1
C[C@H]1[C@@H]2CC3=C(C=C(O)C=C3)[C@@]1(C)CCN2CCC1=CC=CC=C1
C[C@@H]1[C@H]2CC3=C(C=C(O)C=C3)[C@@]1(C)CCN2CCC1=CC=CC=C1
C[C@H]1[C@H]2CC3=C(C=C(O)C=C3)[C@@]1(C)CCN2CCC1=CC=CC=C1

1       (1R,9S,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
2       (1R,9S,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
3       (1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
4       (1R,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
5       (1S,9S,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
6       (1S,9S,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
7       (1S,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
8       (1S,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol

Some of them are probably not possible in 3D.

SMI23D converts all of the 8molecules either into

(1R,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
(1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol

id      Preferred IUPAC Name
1       (1R,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
2       (1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
3       (1R,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
4       (1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
5       (1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
6       (1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
7       (1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
8       (1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol

 

The whole calulation with SMI23D takes one or two seconds, MArvin takes either minutes or it does not come up  woth one, but is says: Calculation failed

 

No possible solution with an output would be correct, or one has to say yes, please allow a change in

configuration. So actually saying failed is not the correct output when all optimizations were checked,

but I am not sure if there shold be a correct solution or a more practial approach to this solution, instead of

nothing  or failed.

Cheers

Tobias

Hi,

I am not sure whats going on,

both molconvert and cxcalc fail with conformer generation and also MarvinView,

but using the Clean 3D fast built (and only fast built works) MView can creat conformers.

Thats very weird. At least conformers a starting points.

See attached file.

Original 8 stereo smiles (8-stereos.smi)

C[C@@H]1[C@@H]2CC3=C(C=C(O)C=C3)[C@]1(C)CCN2CCC1=CC=CC=C1
C[C@H]1[C@@H]2CC3=C(C=C(O)C=C3)[C@]1(C)CCN2CCC1=CC=CC=C1
C[C@@H]1[C@H]2CC3=C(C=C(O)C=C3)[C@]1(C)CCN2CCC1=CC=CC=C1
C[C@H]1[C@H]2CC3=C(C=C(O)C=C3)[C@]1(C)CCN2CCC1=CC=CC=C1
C[C@@H]1[C@@H]2CC3=C(C=C(O)C=C3)[C@@]1(C)CCN2CCC1=CC=CC=C1
C[C@H]1[C@@H]2CC3=C(C=C(O)C=C3)[C@@]1(C)CCN2CCC1=CC=CC=C1
C[C@@H]1[C@H]2CC3=C(C=C(O)C=C3)[C@@]1(C)CCN2CCC1=CC=CC=C1
C[C@H]1[C@H]2CC3=C(C=C(O)C=C3)[C@@]1(C)CCN2CCC1=CC=CC=C1



Z:\smi23d>cxcalc name 8-stereos.smi
id      Preferred IUPAC Name
1       (1R,9S,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
2       (1R,9S,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
3       (1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
4       (1R,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
5       (1S,9S,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
6       (1S,9S,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
7       (1S,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
8       (1S,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol

cxcalc name MARVINView-CLEAN-3D-simple.sdf
id      Preferred IUPAC Name
1       (1S,9S,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
2       (1R,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
3       (1R,9R,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
4       (1R,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
5       (1S,9S,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
6       (1S,9S,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
7       (1S,9S,13S)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol
8       (1R,9R,13R)-1,13-dimethyl-10-(2-phenylethyl)-10-azatricyclo[7.3.1.0^{2,7}]trideca-2(7),3,5-trien-4-ol

So the fast built option will change the absolut stereo configuration.

So the quick way for cleaning a large number of 2D SMILES or 2D SD files,

0) make sure its a unique conformer (R/S) (Z/E) you need or you have

1) generate unique name or unique inchiKey

2) generate 3D conformers

3) check original stereo name or inchiKey

What I see here is that the software could somehow inform the user that there was a change in configuration, at least in the 2D to 3D conversion step. Or the policy is its totally up to the useer, but then the user also has to be warned that the results are not the ones he actually wants or anticipates.

See

MARVINView-CLEAN-3D-simple.sdf

Tobias

Dear Tobias,

Sorry for the late answer and thanks for the very detailed report.

 get an error when I try to generate stereoisomers the first SMILES with Marvin 5.2.2. and 3D check on.

This seems to be a bug, we will investigate it.

We made improvements in Clean3D stereo criteria handling; these modifications will be included in the forthcomming 5.3 release. This version will be able to generate coordinatess for 4 of the 8 stereo isomers (see attached file). The others are seem to be sterically unfeasible.

About stereo centers: Clean3D tries to generate 3D coordinates for the given structure, including RS/EZ specifications. If a stereo center's chirality is specified then should stick to it; if not, the process should choose a suitable configuration. I believe that structures with different stereo specifications should be treated as different molecules (like having diferent connectivity): the default behavior should not change it. 

(Fast cleaning method tries to generate 3D coordinates even if the stereo requrements are not met during the building process.  So it can return "invalid" configuration. This can happen even if the given configuration otherwise is valid.)

Currently we do not have fast, exact solution for checking/fixing stereo configuration - this seems to be a nontrivial problem -, so  stereoisomers plugin does this by trying to generate 3D structures only. (This may change in the future.)

Regards,

Gabor