Do I need Evaluator for reactions containing rules?

My final set of questions:





How is the evaluator used in the context of including calculator plug-ins in chemical terms? If I want to use the reactor API to run my reactions in a Java wrapper class do I need to include the Evaluator API class and construct an instance of Evaluator or does the react API call evaluator when it recognises a reference to a plug-in in the appended chemical terms of a reaction string?





Secondly, it appears that I can calculate my plug-ins without the need for evaluator: examples, being





http://www.jchem.com/doc/api/chemaxon/marvin/calculations/HBDAPlugin.html





and Slide 9 of the recent Calculator Plug-ins presentation.





I guess I am struggling a bit to see how evaluator fits in to determining the calculator plug-ins in both a reaction and stand alone context - is it an essential component in either context?





Where can I see a list of prices for the calculator plug-ins please?


Would also be useful to see a price for a single react licence.

Please find attached two versions of the Wittig reaction, one to produce the E isomer and one for the Z isomer.





I run the reactions for the two input sets attached using combinatorial mode and get no output (and no errors) - could you please see where I am going wrong in my reaction definitions please!? I don't think this is a chemical terms related issue.





Many thanks,


Daniel.

Year/perpetual prices for:





TPSA


PKa


Log P


Huckel


HBDA





would be good to start with.

Our CBO will contact you soon regarding the prices.

inhibox wrote:

How is the evaluator used in the context of including calculator plug-ins in chemical terms? If I want to use the reactor API to run my reactions in a Java wrapper class do I need to include the Evaluator API class and construct an instance of Evaluator or does the react API call evaluator when it recognises a reference to a plug-in in the appended chemical terms of a reaction string?

No, you do not need to call Evalutor explicitly, it is accessed by Reactor automatically if the reaction has rules.


Rules are typically specified in Molecule properties (REACTIVITY, SELECTIVITY, TOLERANCE) saved in RDF/MRV tags, or else they can be specified in the





setReaction() method (there are more overloaded versions setting different rules). I think the best is to include the rules in properties / tags but unfortunately msketch is currently not capable of setting these proeprties (although it keeps them if you edit/save your reaction). To set these properties, you can either





- manually edit the reaction file (I attach some sample files), or


- you can use our


Reactor JSP application (set the rules in the text areas and save the reaction by pressing "Save reaction")





If you choose the latter then I would very much appreciate your feedback about the JSP interface.

Quote:

Secondly, it appears that I can calculate my plug-ins without the need for evaluator: examples, being http://www.jchem.com/doc/api/chemaxon/marvin/calculations/HBDAPlugin.html and Slide 9 of the recent Calculator Plug-ins presentation. I guess I am struggling a bit to see how evaluator fits in to determining the calculator plug-ins in both a reaction and stand alone context - is it an essential component in either context?

Yes, you can access the plugins by their specific API directly from a JAVA program and you can also use the


cxcalc


command line tool for plugin calculations.


To evaluate chemical expressions directly you can use the


evaluate


command line tool or the Evaluator API from a JAVA program. This can be useful if you want to post-filter your products by drug-likeness or other filtering expression. However, reaction rules (reactivity, selectivity) are best to be included in the reaction because these rules are then automatically evaluated in the reaction context before returning the products. The reaction context enables you to refer to reactants and products (e.g. reactant(1), product(0)) and reactant and product atoms matching specified map numbers in the reaction equation (e.g. ratom(1), patom(2)), which is not possible outside the reaction processing mechanism.

Quote:

Where can I see a list of prices for the calculator plug-ins please? Would also be useful to see a price for a single react licence.

License details:


http://www.chemaxon.com/License_Details.html





Our CBO will contact you soon regarding the prices.

Quote:

Please find attached two versions of the Wittig reaction, one to produce the E isomer and one for the Z isomer. I run the reactions for the two input sets attached using combinatorial mode and get no output (and no errors) - could you please see where I am going wrong in my reaction definitions please!? I don't think this is a chemical terms related issue.

I was able to run the Wittig reactions, but by default, all atoms having changing bonds must be mapped. So, please add a map number (6) to the oxygene which eliminates during the reaction, and also draw a mapped (7) explicit hydrogen eliminating from carbon 4.

Code:

[#6:2][C:1]=[O:6].[H:7][C:4]([#6:3])[#9,#17,#35,#53:5]>>[#6:2]\[C:1]=[C:4]/[#6:3]

Since your reaction equation contains a carbonyl with an alpha carbon and all of your oxo reagents are ketons, there will be no cis/trans selectivity for the reactions. Your carbonyl reactant will match both alpha carbons of your ketons, thus Reactor will generate both cis/trans isomers as products.





(The double bond stereo part has been significantly improved in the upcoming JChem release. From that new version all double stereo configuration will be determined from the drawn 2D geometry and the diastereomeric double bonds can be marked with wiggly substituents.)

inhibox wrote:

I guess I am struggling a bit to see how evaluator fits in to determining the calculator plug-ins in both a reaction and stand alone context - is it an essential component in either context?

All chemical expression evaluation tools (Reactor, Evaluator, Search) access the Evaluator


class to actually evaluate the expression on the given inputs.


On the other hand, cxcalc directly accesses the plugins and is not capable of complex expression evaluation. For example,

Code:

cxcalc logp "OC1=CNC=C1Cl" "OC(=O)CCC1=CNC=C1Cl" id      logP 1       1.29 2       1.60

will calculate the logP values for two input molecules. You can do the same with Evaluator:

Code:

evaluate -e "logp()" "OC1=CNC=C1Cl" "OC(=O)CCC1=CNC=C1Cl" 1.29 1.6

But: to test which of them has bigger logP than a certain limit, you should use Evaluator:

Code:

evaluate -e "logp() > 1.5" "OC1=CNC=C1Cl" "OC(=O)CCC1=CNC=C1Cl" 0 1

where 0 means FALSE and 1 means TRUE.

Gyuri wrote:

(The double bond stereo part has been significantly improved in the upcoming JChem release. From that new version all double stereo configuration will be determined from the drawn 2D geometry and the diastereomeric double bonds can be marked with wiggly substituents.)

The wiggly-bond version of your reaction in SMARTS:

Code:

[C:1]([#6:2])=[O:6].[#6:3][C:4]([#9,#17,#35,#53:5])[H:7]>>[C:1]([#6:2])=[C:4][#6:3]

The difference: no '/' and '\' signs on the product side.


I also attach this in MRV format.


From the next major JChem release, the products will also be marked with wiggly-bonds.

I'm sorry to report that I still cannot get the Wittig reaction working.


I have tried your SMARTS example saved as mrv and rxn and yet no output is apparent.





May I suggest that the O and H are not mapped in the SMARTS that you have provided? Presumably the L requires mapping also?





I have tried quite a few variations - maybe you could upload a rxn version and some inputs that you know work.





Sorry + Thanks,


Daniel.

1. Goto this page:


http://www.jchem.com/examples/reactor/jsp/index.jsp





2. Paste the SMARTS or load the attached wittig.rxn file in the reaction window.





3. Press the "Set Reactants" button at the bottom of the page.





4. Add your Ketone.smi file as first reactant, and HaloAlkane.smi file as the second reactant.





5. Select "combinatorial" mode and "reaction" type options at the bottom of the page.





6. Press the React button.





18 reactions are generated.

That's a great way to do it, which I will use to test from now on!





However in this case I have completed and get the message "no products created" using the SMARTS you suggested for Wittig reaction - I take it that you do see some products at this stage?

OK - it works fine with the wittig.rxn file you provided.


Great thanks for your help.


Will now try and run with react call at this end.

Strangely, a similar call to react tool produces no output though?





react C:\reactor\Wittig\HaloAlkane.smi C:\reactor\Wittig\Ketone.smi -r C:\reactor\wittig\wittig.rxn -o C:\reactor\wittig\WittigOutput.sdf -m comb

Which JChem version do you use?

I was using the react tool which came with 3.0.10?


The rest of my JChem set up is 2.3 and I am due to upgrade next week!


(An indication of when 3.1 is out would be useful)


Also I only have a trial license key so this may further complicate.

inhibox wrote:

Strangely, a similar call to react tool produces no output though? react C:\reactor\Wittig\HaloAlkane.smi C:\reactor\Wittig\Ketone.smi -r C:\reactor\wittig\wittig.rxn -o C:\reactor\wittig\WittigOutput.sdf -m comb

You have changed the reactant order!


Try:

Code:

react C:\reactor\Wittig\Ketone.smi C:\reactor\Wittig\HaloAlkane.smi -r C:\reactor\wittig\wittig.rxn -f sdf -o C:\reactor\wittig\WittigOutput.sdf -m comb

Note, that you should specify the output format in parameter -f if the format is not SMILES.

Sorry my mistake - it works perfect!





All in all, my questions are fully answered, so thank you very much for your great support. We will be in touch at the latest, by the end of next week with respect to purchase of relevant licesnses.





It would be useful to have all the price details by the end of the day - or at least available to view by Monday?


i.e. Prices for...


1 * full react license (not the react pro for now)


5 * calculator plug - ins mentioned above.