Problem with predicting SMILES

We are currently running JChem 3.2.5 and Marvin 4.1.7. We are having a problem with a pattern that is supposed to predict hydrolysis of a cyclic ester. We have another pattern (bt0024) that predicts ester hydrolysis but it is less stereospecific than the second pattern (bt0288a). The substrate we are testing is below:

Code:

CC1=CC(=O)OC(CC(O)=O)C1(C)C

We would like bt0288 to produce:

Code:

C\C(=C\C([O-])=O)C(C)(C)\C=C\C(O)=O

Here is what we're getting:

Code:

/home/sdml/umbbd/src/java> react -a changing -v -r ../molecules/bt0024.mrv "CC1=CC(=O)OC(CC(O)=O)C1(C)C" CC(=CC([O-])=O)C(C)(C)C(O)CC(O)=O CC(=CC([O-])=O)C(C)(C)C(O)CC(O)=O Total running time (ms) : 986 Reaction setting (ms)   : 794 Reactant setting (ms)   : 0 Reaction processing (ms): 192 /home/sdml/umbbd/src/java> react -a changing -v -r ../molecules/bt0288a.mrv "CC1=CC(=O)OC(CC(O)=O)C1(C)C" Total running time (ms) : 521 Reaction setting (ms)   : 350 Reactant setting (ms)   : 2 Reaction processing (ms): 169 /home/sdml/umbbd/src/java>

I've attached both patterns below. Does anyone have an idea as to why bt0288 does not (and bt0024 does) seem to recognize the SMILES for this structure?

Hi,





Try the attached (bt0288a_mod.mrv) reaction. I have modified the double bond of the reactant to "any" bond, because this caused the error (the Reactor engine was searching for a cis double bond in input reactants).





This behavior will be corrected in JChem 5.0.





Zsolt