Inability to predict SMILES strings

We are having a problem with one of our reactions refusing to predict SMILES strings. No matter what substrate or product we enter into Marvin Sketch, we always get the error message "unable to produce predicted product SMILES". We have reported this problem in the past and this reaction has not worked in jchem 3.1.7, 3.2.1, or our current version, 3.2.3. We are currently using Marvin Sketch 4.1.5.

Code:

react -a changing -v -r ../molecules/bt0085.mrv "O[C@H]1C=CC2=C([C@H]1O)c3ccccc3C2=O" Total running time (ms) : 405 Reaction setting (ms)   : 251 Reactant setting (ms)   : 2 Reaction processing (ms): 152

We get running times but no products are predicted. I have also attached the reaction as a .mrv file (called bt0085). Any ideas on why this rule is a problem? Thanks in advance.

Hi,


with reactor 3.2.1 command line/GUI and the reaction files - this is a feature (not a bug - edited), because the arrow in Msketch is red. (would be a nice feature to show the source of errors :-) Now, even if you add water its still reporting zero reactions in the GUI (so its a bug, not a feature).





But you can use this one instead by expanding everything to SMILES, it works:


react -r "[OH:3][C@H:2]1C=CC2=C([C@H:1]1[OH:4])C3=C(C=CC=C3)C2=O>>[OH:4][C:1]1=[CH:2]C=CC2=C1C3=C(C=CC=C3)C2=O" "O[C@H]1C=CC2=C([C@H]1O)c3ccccc3C2=O"

Code:

C:\chemistry\jchem\tk\react\test>react -v -r "[OH:3][C@H:2]1C=CC2=C([C@H:1]1[OH:4])C3=C(C=CC=C3)C2=O>>[OH:4][C:1]1=[CH:2]C=CC2=C1C3=C(C=CC=C3)C2=O"  "O[C@H]1C=CC2=C([C@H]1O)c3ccccc3C2=O" OC1=CC=CC2=C1c3c:cc:cc3C2=O Total running time (ms) : 732 Reaction setting (ms)   : 311 Reactant setting (ms)   : 10 Reaction processing (ms): 411 C:\chemistry\jchem\tk\react\test>

Tobias

Hi,

Rlong wrote:

We are having a problem with one of our reactions refusing to predict SMILES strings. No matter what substrate or product we enter into Marvin Sketch, we always get the error message "unable to produce predicted product SMILES". We have reported this problem in the past and this reaction has not worked in jchem 3.1.7, 3.2.1, or our current version, 3.2.3. We are currently using Marvin Sketch 4.1.5. Code: react -a changing -v -r ../molecules/bt0085.mrv "O[C@H]1C=CC2=C([C@H]1O)c3ccccc3C2=O" Total running time (ms) : 405 Reaction setting (ms)   : 251 Reactant setting (ms)   : 2 Reaction processing (ms): 152 We get running times but no products are predicted. I have also attached the reaction as a .mrv file (called bt0085). Any ideas on why this rule is a problem? Thanks in advance.

This is a bug. There is no problem with the reaction.

TobiasKind wrote:

But you can use this one instead by expanding everything to SMILES, it works: react -r "[OH:3][C@H:2]1C=CC2=C([C@H:1]1[OH:4])C3=C(C=CC=C3)C2=O>>[OH:4][C:1]1=[CH:2]C=CC2=C1C3=C(C=CC=C3)C2=O" "O[C@H]1C=CC2=C([C@H]1O)c3ccccc3C2=O"

Yes, this works.





You can also specify the "-z" (transform mode) command line switch to make it work:

Code:

$ react -a changing -z -r bt0085.mrv "O[C@H]1C=CC2=C([C@H]1O)c3ccccc3C2=O" OC1=CC=CC2=C1c3ccccc3C2=O

Please use one of these methods as workaround until we fix this issue.





Note: "-z" can be used only with 1-reactant -> 1-product reactions.

TobiasKind wrote:

with reactor 3.2.1 command line/GUI and the reaction files - this is a feature (not a bug - edited), because the arrow in Msketch is red. (would be a nice feature to show the source of errors :-) Now, even if you add water its still reporting zero reactions in the GUI (so its a bug, not a feature).

The red arrow in msketch denotes that the reaction is non-stochiometric, but Reactor doesn't mind it.





Best regards,


Zsolt

Here is a workaround for the problem (which I think is related to predicting cis/trans configuration of double bonds in an aliphatic ring). The rule is more general than the original but at least it works.