Pka (batch mode and individual)

User 5b7b568917

08-10-2007 19:35:28

Hello ,





I have a set of molecules for which I am calculating Pka for.





I used the batch process which gave me the result with four colums and the atom number where the ionization occurs. My question here is how will I know which atom number corresponds to which atom.





Next , i used the marvin sketch and drew individual molecules to find the pka's(the same I used for batch process). For some molecules(for every 2 in 10 molecules), I see that the value of pka calculated is no way close to the ones that are calculated by batch.





Could you please tell me which one is correct and how i need to figure this out.





Thanks,


Usha

ChemAxon e08c317633

09-10-2007 11:15:16

Hi,
ucheemak wrote:
I have a set of molecules for which I am calculating Pka for.





I used the batch process which gave me the result with four colums and the atom number where the ionization occurs. My question here is how will I know which atom number corresponds to which atom.
MarvinSketch and MarvinView can show atom numbers. To display atom numbers choose "View->Misc->Atom Numbers" from menu.





Example:


Code:
$ cxcalc pka "NCC(O)=O"


id      apKa1   apKa2   bpKa1   bpKa2   atoms


1       2.31            9.24            4,1





The atom with index 4 has acidic pKa of 2.31, and the atom with index 1 has basic pKa of 9.24. I have attached the image of the molecule with atom numbers.
Quote:
Next , i used the marvin sketch and drew individual molecules to find the pka's(the same I used for batch process). For some molecules(for every 2 in 10 molecules), I see that the value of pka calculated is no way close to the ones that are calculated by batch.





Could you please tell me which one is correct and how i need to figure this out.
Can you post a few molecules which gave different results in cxcalc and in MarvinSketch?





Regards,


Zsolt

User 5b7b568917

09-10-2007 13:36:37

Thanks for your reply. But since I am using batch process to process the molecules (~ 1500) , is there a way that the atom name that it corresponds to will be displayed along with the atom number. Or do I need to individually load the molecule into the marvin sketch and see which atom the correspondinf pka value belongs to.





The following are some of the SMILES that gave me different values for pka using cxcalc and marvin sketch.











[H]OC([H])([H])[C@@]1([H])O[C@@]([H])(N2C([H])=NC3=C2N=C([H])N=C3N([H])[H])C([H])([H])[C@]1([H])OP(=O)(O[H])O[H]





[H]OC(=O)[C@@]([H])(N([H])C([H])([H])C1=C([H])C(=C([H])C([H])=C1O[H])C(\N([H])[H])=N\[H])C([H])([H])[H]





[H]OC(=O)[C@@]([H])(N([H])C([H])([H])C1=C([H])C([H])=C(C([H])=C1O[H])C(\N([H])[H])=N\[H])C([H])([H])[H]








Thanks,


Usha

User 870ab5b546

10-10-2007 02:52:47

I may be wrong, but it appears to me that atom numbers are assigned to the atoms in a SMILES string simply in the order that they appear in the string. In the example, NCC(O)=O, the N atom is atom 1 and the O atom with no double bonds is atom 4. So you can easily write a little program that determines which atoms have which numbers, and hence have which pK values.

ChemAxon 25dcd765a3

10-10-2007 11:42:16

Yes.


Bob you are right. It is exactly as you have written:
Quote:



atom numbers are assigned to the atoms in a SMILES string simply in the order that they appear in the string


ChemAxon e08c317633

10-10-2007 12:22:03

ucheemak wrote:
Thanks for your reply. But since I am using batch process to process the molecules (~ 1500) , is there a way that the atom name that it corresponds to will be displayed along with the atom number. Or do I need to individually load the molecule into the marvin sketch and see which atom the correspondinf pka value belongs to.
You can also use molconvert to generate images of the molecules with atom numbers. It can accept any input format we support, and works in batch mode.





See also: image export in Marvin





Example:


Code:
molconvert png:anum mol1.sdf -o mol1.png
Quote:
The following are some of the SMILES that gave me different values for pka using cxcalc and marvin sketch.





[H]OC([H])([H])[C@@]1([H])O[C@@]([H])(N2C([H])=NC3=C2N=C([H])N=C3N([H])[H])C([H])([H])[C@]1([H])OP(=O)(O[H])O[H]





[H]OC(=O)[C@@]([H])(N([H])C([H])([H])C1=C([H])C(=C([H])C([H])=C1O[H])C(\N([H])[H])=N\[H])C([H])([H])[H]





[H]OC(=O)[C@@]([H])(N([H])C([H])([H])C1=C([H])C([H])=C(C([H])=C1O[H])C(\N([H])[H])=N\[H])C([H])([H])[H]


Which version of Marvin are you using? What are the results you get with MarvinSketch, and what are the results with cxcalc? We tried to reproduce the error, but we got identical results.





Note: cxcalc with default settings displays only the first two strongest acidic and basic pKa-s. You can use "-a" and "-b" options to display more acidic/basic pKa-s.





Code:
cxcalc pka -a 4 -b 4 "[H]OC([H])([H])[C@@]1([H])O[C@@]([H])(N2C([H])=NC3=C2N=C([H])N=C3N([H])[H])C([H])([H])[C@]1([H])OP(=O)(O[H])O[H]"


id      apKa1   apKa2   apKa3   apKa4   bpKa1   bpKa2   bpKa3   bpKa4   atoms


1       1.74    6.74    14.77           3.94    -2.90   -3.36   -4.68   33,35,2,20,2,14,8








Regards,


Zsolt

User 5b7b568917

10-10-2007 14:05:57

Hai,





Please find the following smiles, that gave me the pka's using cxcalc.





Smiles : P(OC1CC(OC1CO)n1c2ncnc(N)c2nc1)(O)(O)=O


apka1 1.74


apka2 6.74


bpka1 3.94


bpka2 -2.9


atoms 3,4,19,5








I am attaching the file that gave me pka using Marvin sketch.





Please see the difference. Even when you say that cxcalc gives two acidic and two basic values, these two values are not present when I used marvin sketch. It gave me a bunch of values.





Thanks,


Usha








ChemAxon e08c317633

10-10-2007 15:15:44

Hi,





You have used different pKa options in MarvinSketch and in cxcalc. In MarvinSketch you have used "micro" mode, while in cxcalc you have used "macro" mode ("macro" mode is the default).





Set the "Mode" parameter to "macro" on pKa options panel to get the results you got in cxcalc.





If you want to perform "micro" mode pKa calculation with cxclac, you can do it like this:





Code:
cxcalc pka -m micro "P(OC1CC(OC1CO)n1c2ncnc(N)c2nc1)(O)(O)=O"


id      apKa1   apKa2   bpKa1   bpKa2   atoms


1       2.05    2.05    3.94    1.60    20,21,14,12






Regards,


Zsolt