another stereochemistry issue

Reaction definition:

<?xml version="1.0" ?>
<MDocument>
  <MChemicalStruct>
    <reaction x1="-2.3919096628986525" y1="3.519216030990243" x2="3.5414076794319564" y2="3.271128347664041">
      <propertyList>
        <property dictRef="NAME" title="NAME">
          <scalar><![CDATA[epoxides with (-)CN]]></scalar>
        </property>
        <property dictRef="SELECTIVITY" title="SELECTIVITY">
          <scalar>-stericHindrance(ratom(1))</scalar>
        </property>
      </propertyList>
      <reactantList>
        <molecule molID="m1">
          <atomArray
              atomID="a1 a2 a3 a4"
              elementType="C O H C"
              mrvMap="1 3 0 0"
              reactionStereo="Inv 0 0 0"
              x2="-6.881875038146973 -5.341875038146973 -7.280456367604854 -6.111875038146972"
              y2="4.042500257492065 4.042500257492065 5.530026029977231 2.7088211356640297"
              />
          <bondArray>
            <bond atomRefs2="a1 a4" order="1" />
            <bond atomRefs2="a1 a3" order="1" />
            <bond atomRefs2="a1 a2" order="1" />
            <bond atomRefs2="a4 a2" order="1" />
          </bondArray>
        </molecule>
      </reactantList>
      <productList>
        <molecule molID="m2">
          <atomArray
              atomID="a1 a2 a3 a4 a5 a6 a7"
              elementType="C C H C O H N"
              mrvMap="2 1 0 0 3 0 0"
              x2="7.56771259790619 7.601281909793624 6.860536746513022 9.088807682278789 9.088807682278789 10.422486804106825 7.56771259790619"
              y2="2.4937202768337343 4.033354357159123 5.383501239196915 4.431935686617004 5.971935686617004 6.741935686617004 0.9537202768337343"
              />
          <bondArray>
            <bond atomRefs2="a1 a7" order="3" />
            <bond atomRefs2="a1 a2" order="1" />
            <bond atomRefs2="a2 a4" order="1" />
            <bond atomRefs2="a2 a3" order="1" />
            <bond atomRefs2="a4 a5" order="1" />
            <bond atomRefs2="a5 a6" order="1" />
          </bondArray>
        </molecule>
      </productList>
    </reaction>
  </MChemicalStruct>
</MDocument>

If I submit [H][C@@]12CCCC[C@]1([H])O2, Reactor produces only one product, [H]O[C@]1([H])CCCC[C@@]1([H])C#N.  It fails to produce the other enantiomer.  This is a problem because it makes it seems as if this symmetry-breaking reaction is asymmetric, and it's not.

In general, meso starting material -> chiral product should cause Reactor to give both enantiomers.  

It works for me. Reactor 5.4.1.1 generates two products:

react -r StereochemistryIssue2.mrv "[H][C@@]12CCCC[C@]1([H])O2"
[H][C@]1(O)CCCC[C@]1([H])C#N
[H][C@@]1(O)CCCC[C@@]1([H])C#N

Zsolt

Well here's my code (MAX_REACT is 6):

    private final Reactor reactor = new Reactor();

    ...

            moleculeOfRxn = new MolHandler(singleRxnDef.trim()).getMolecule();

            rxnMol = RxnMolecule.getReaction(moleculeOfRxn);

            if (rxnMol != null) reactor.setReaction(rxnMol);

            ...

            caller.debugPrint("Permutation ", perm, " of ", numPerms,

                    ": ", starters);

            reactor.setReactants(starters);

            boolean anySMsAreChiralNonracemic = false;

            try {

                anySMsAreChiralNonracemic =

                        StereoFunctions.anyAreChiralNonracemic(starters);

            } catch (MolFileException e) {

                Utils.alwaysPrint(SELF + "caught MolFileException "

                        + "trying to get whether any of ",

                        starters, " are chiral.");

            } // try

            // Reactor will give further products with each call

            for (int reactCt = 0; reactCt < MAX_REACT; reactCt++) {

                reactorProds = reactor.react();

                if (Utils.isEmpty(reactorProds)) {

                    caller.debugPrint("Product set ", reactCt + 1,

                            " from Reactor for permutation ", perm,

                            " is empty; breaking.");

                    break;

                } // if there are no products

                gotProducts = true;

                caller.debugPrint("Product set ", reactCt + 1,

                        " obtained from Reactor for permutation ", perm, ".");

                final int numReactorProds = reactorProds.length;

                convert0DTo2D(reactorProds);

                if (numPriorRxnProds > 0) {

                    ...

                } else caller.debugPrint("Initial substrates reacted to give ",

                        numReactorProds, " product(s): ", reactorProds);

                implicitizeHAtoms(reactorProds);

                caller.debugPrint("After implicitizing H atoms, ",

                        "Reactor products are: ", reactorProds);

                // add info on relative yields, whether product

                // should be racemized

                int prodNum = 0;

                for (Molecule reactorProd : reactorProds) {

                    ...

                } // for each reactor prod

                // Add each product to this loop's product list unless 

                // it's already in the list or is a starting material.

                addToRxnLoopProducts(reactorProds, thisRxnLoopProducts);

                if (generatedSMStereo != NO_SM_RACEMIZED) return;

            } // while still getting products from reactor

And here's my log:

Permutation 1 of 1: [C1CC[C@@H]2O[C@@H]2C1]

Product set 1 obtained from Reactor for permutation 1.

Initial substrates reacted to give 1 product(s): [[H]O[C@@H]1CCCC[C@H]1C#N]

After implicitizing H atoms, Reactor products are: [O[C@@H]1CCCC[C@H]1C#N]

Setting reactor index property for reactor product 1, O[C@@H]1CCCC[C@H]1C#N, to 1

Setting racemize property for reactor product 1, O[C@@H]1CCCC[C@H]1C#N, to true.

Product O[C@@H]1CCCC[C@H]1C#N not already in initial substrate or product lists, [C1CC[C@@H]2O[C@@H]2C1], [], []; adding it to thisRxnLoopProducts.

Product set 2 from Reactor for permutation 1 is empty; breaking.

Also:

bob@epoch-virtual:web$ react -r tempfiles/StereochemistryIssue2.mrv "[H][C@@]12CCCC[C@]1([H])O2"

[H]O[C@]1([H])CCCC[C@@]1([H])C#N

bob@epoch-virtual:web$ 

So I am observing a different result from you.  Why?  Again, I am using JChem 5.4.1.  Same result with JChem 5.4.1.1.

I guess this is a configuration issue. I get the same 2 output molecules with JChem 5.4.1. and JChem 5.4.1.1. too (see my previous post).

Bob, could you post here the output of

echo $CLASSPATH

which react

and

react -h | grep Reactor 

commands?

Can you execute the same command on a different computer with a fresh JChem 5.4.1.1 install, and post the result here?

Zsolt

With JChem 5.4.1 on our server:

bob@epoch-virtual:~$ echo $CLASSPATH

/home/aceorg/aceorg/DEV-3.1/ace/build:/usr/share/java:/home/aceorg/aceorg/jchem5.4.1/lib:/home/aceorg/aceorg/otherJar/ojdbc14.jar

bob@epoch-virtual:~$ which react

/home/aceorg/aceorg/jchem5.4.1/bin/react

bob@epoch-virtual:~$ react -h | grep Reactor

Reactor 5.4.1.1, (C) 1999-2011 ChemAxon Ltd.

bob@epoch-virtual:~$ 

Then I fresh-installed JChem 5.4.1.1 and changed my PATH and CLASSPATH variables.

bob@epoch-virtual:web$ react -r tempfiles/StereochemistryIssue2.mrv "[H][C@@]12CCCC[C@]1([H])O2"

[H]O[C@]1([H])CCCC[C@@]1([H])C#N

bob@epoch-virtual:web$ echo $CLASSPATH

/home/aceorg/aceorg/DEV-3.1/ace/build:/usr/share/java:/home/aceorg/aceorg/jchem5.4.1.1/lib:/home/aceorg/aceorg/otherJar/ojdbc14.jar

bob@epoch-virtual:web$ which react

/home/aceorg/aceorg/jchem5.4.1.1/bin/react

bob@epoch-virtual:web$ react -h | grep Reactor

Reactor 5.4.1.1, (C) 1999-2011 ChemAxon Ltd.

bob@epoch-virtual:web$ 

Then I fresh-installed JChem 5.4.1.1 on my desktop.

Grossman-CP-339-iMac:~ Bob$ react -r Desktop/StereochemistryIssue2.mrv "[H][C@@]12CCCC[C@]1([H])O2"

[H]O[C@]1([H])CCCC[C@@]1([H])C#N

Grossman-CP-339-iMac:~ Bob$ echo $CLASSPATH

/usr/share/java:Applications/ChemAxon/JChem5.4.1.1/lib

Grossman-CP-339-iMac:~ Bob$ which react

/Applications/ChemAxon/JChem5.4.1.1/bin/react

Grossman-CP-339-iMac:~ Bob$ react -h | grep Reactor

Reactor 5.4.1.1, (C) 1999-2011 ChemAxon Ltd.

Grossman-CP-339-iMac:~ Bob$ 

Same results every time.

Bob, we used different reaction definitions, that's why we got different results. In your reaction definition (first post) there is a selectivity rule: -stericHindrance(ratom(1)), while in my reaction definition (StereochemistryIssue2.mrv, attached in second post) there is no selectivity rule.

Try this, it will ignore selectivity rule, and will generate two products with both reaction definitions:

react -r StereochemistryIssue2.mrv -n s "[H][C@@]12CCCC[C@]1([H])O2"
[H][C@]1(O)CCCC[C@]1([H])C#N
[H][C@@]1(O)CCCC[C@@]1([H])C#N

With the selectivity rule the reaction generates one product, because the steric hindrance difference on ratom(1) is bigger than 0.0001 (default tolerance):

Steric hindrance on ratom(1) in first reaction:  1.000775
Steric hindrance on ratom(1) in second reaction:  0.998256

The reactant is symmetric, so steric hindrances should be equal on these atoms. They are different because the 3D generator does not generate absolutely symmetric conformers.

So the proper solution in this case  is to increase the tolerance.

Zsolt

Thanks for sleuthing this out and for proposing a workaround.  The fact that the 3D generator does not account for the symmetry of the compound is a deficiency that I encourage you to remedy in the future.  It is completely counterintuitive to have to include a tolerance to account for symmetrical molecules.  At the very least, you need to document the need to do this on your page on writing reaction definition rules.

We will add this to the documentation of steric hindrance.