suggestion for calculation of membrane logP and membrane pKa

Hello ChemAxon





I have been using Marvin 5.0 on my Mac for about 2 months now and I am very satisfied with the product. I have only had a few glitches so far (like difficulty obtaining a good prediction of pKa for NH groups; try the molecules FCCP (exp. pKa = 6.2) and CCCP (exp. pKa = 6.0) and you'll see what I mean). And I wish the GUI was a little better adapted to my Mac (if anyone is listening, Please allow windows to close with apple-C and please make OK the default choice in tool windows so that I can use my keyboard instead of my mouse, and please make the Open and Save file browsers OSX-compliant).





Today's question: Can the pKa plugin be used with user-supplied pKa values instead of structure derived pKa values? This would be a useful little implementation to calculate speciation (distribution of ionized species). Perhaps a little input window with up to 5 or 6 pKa values?





The reason I am asking is that I would like to correct your bulk pKa values to reflect lipid bilayer membrane pKa values.





What would be tremendously useful for me would be the ability to combine the logP plug-in with the pKa plug-in in a novel way: You could include a correction of octonol logP values to membrane log P (logPmem) values (there are solid equations in the literature)(this would be useful to many users), and then use the difference in log Pmem between neutral and ionized species (delta logPmem) to adjust bulk pKa to membrane pKa (pKa mem; pKa mem = pKa - delta logPmem). The delta would have to be calculated for each microspecies of course. Once the new pKa values are obtained, then the speciation calculation is performed to estimate the true ionization behaviour of the ionizable molecule in a biological membrane.





This is very tedious to do by hand. This type of integration between the logP and the pKa plugins would save me countless hours.





Thanks for the support!





Cheers


Louis Martineau

louism wrote:

And I wish the GUI was a little better adapted to my Mac (if anyone is listening, Please allow windows to close with apple-C and please make OK the default choice in tool windows so that I can use my keyboard instead of my mouse, and please make the Open and Save file browsers OSX-compliant).

Hello Louis,





thank you for the GUI related notes, we will investigate them and return with the result soon.

Quote:

Please allow windows to close with apple-C

This one is easy with configuring the shortcuts. The details can be found here: http://www.chemaxon.com/marvin/help/sketch/gui/customization.html#shorcuts


If you do not succeed with the document, please tell us, and we will try to make it better.





My colleagues will answer to your other questions soon.





Kind regards,


Judit

Hi Judit





Thank you for addressing so quickly my little comments on the Mac GUI.





Note that I meant to say Apple-W, not Apple-C. Sorry.





In the world of Macs, Apple-W closes the current window. In Marvin 5.0, Apple-W quits the application. It would be very useful to be able to close results windows generated by the plug-ins using the keyboard command. Thanks. And like I said, if the default button was OK instead of Cancel, I could launch plug-in calculations by pressing Return on my keyboard instead of using my mouse.





Another suggestion (although it is more of a feature request than a GUI issue):





**Could the logP plug-in have an option for automatic calculation of all possible ionized species? Right now, I have to manually change the charge on each ionizable group and repeat the calculation for each group. This ties in with my earlier suggestion for closer interaction between the logP and the pKa plugins (see original post in this thread)**





And a last very minor point: could you put in an option for not using MarvinSpace when calculating logP? It would speed things up if no 3D rendering is necessary.





Cheers


Louis

Hi,

Quote:

Can the pKa plugin be used with user-supplied pKa values instead of structure derived pKa values? This would be a useful little implementation to calculate speciation (distribution of ionized species). Perhaps a little input window with up to 5 or 6 pKa values?

This type of data input is not available in the pKa plugin and we don't plan to implement it in the near future.

Quote:

...membrane log P (logPmem) values (there are solid equations in the literature)... ...bulk pKa to membrane pKa (pKa mem; pKa mem = pKa - delta logPmem)...

Thank you for this information. I think you need to write your own java program and implement these equations.

Quote:

Could the logP plug-in have an option for automatic calculation of all possible ionized species?

You can generate the ionized species and save them into a file. For example with the next command line the ionized species are generated and saved in the 'msp.sdf' file.

Code:

cxcalc microspeciesdistribution yourMolecule > msp.sdf

After this the logP of the ionized species can be calculated and saved in the file 'result.txt' with the next command.

Code:

cxcalc logP -t logPMicro msp.sdf  > result.txt

Jozsi

Hi Louis,

louism wrote:

And a last very minor point: could you put in an option for not using MarvinSpace when calculating logP? It would speed things up if no 3D rendering is necessary.

There is already an option for this. Just uncheck the "Display in MarvinSpace" checkbox on the logP options panel.





Regards,


Zsolt

Hi Zsolt





You are of course correct. I meant the molecular surface area calculator, not the logP calculator. My mistake.





Hi Jozsi





Of course there are no current plans to implement user-defined pKa values in the speciation calculator. It is a new suggestion for increased versatility to an already great program. Please consider it. I picture input boxes beside each calculated pKa value in the structure diagram. If the source code is accessible, and such user-implemented modifications are encouraged, perhaps I will attempt it myself.





As to my suggestions concerning the prediction of alternative partition coefficients, you should be aware that logP octonol is a very poor indicator of physiological partition, especially when considering ionized molecules. Implementation of more physiological predictors would give Marvin a big advantage over other physchem software packages. Even adding models of absorption based on logPoct values (such as a simple parabolic function with a peak in the +2 to +3 range) would be a valuable and easily implemented option.





Cheers


Louis

louism wrote:

You are of course correct. I meant the molecular surface area calculator, not the logP calculator. My mistake.

Hi,





Molecular surface area plugin uses the 3D structure of the molecule for calculation. If the input is a 3D structure, then it is used for the calculation and the 3D structure generator is not run, otherwise (if the input molecule has no 3D coordinates) the 3D structure generator is run and can not be switched off.





You can use polar surface area plugin (in fact, it is topological polar surface area plugin) if it suits your needs. It is topology based.





Regards,


Zsolt