pKa plugin applicability

Hi all!





I am using the pkaPlugin and/or MajorMicrospeciesPlugin in my programs. When running on a large dataset and checking the results, I was wondering about a few discrepancies, for example guanine:





ChemAxon/MarvinBeans/bin/cxcalc pka "NC1=NC2=C(N=C[NH]2)C(=O)N1"


id apKa1 apKa2 bpKa1 bpKa2 atoms


1 6,56 12,40 4,11 0,66 8,11,3,6





From the experimental values, I would expect a distribution like this:


apka1 = 9.4 (atom 11)


apka2 = 12.6 (atom 8)


apkb1 = 10.8 (atom 6)





So my question is: do you have some criteria for discrimiation: for which molecules does the method work well and for which it does not? Maybe it is documented somewhere and I didn't find it - please give me a small hint in this case.





The previous version of the plugin (Marvin 3.3.3, I think) gave a totally different result. Has there been some change in the algorithm?





I am currently using JChem 3.1.6 / Marvin 4.0.5.





Kai

Dear Kai,








Calculated pKa 6.56 seems to be too low. I will check what is the problem.








I think 10.8 is pKb value. 14.0 - 10.8 ==> pKa= 3.2 .








Our pKa calculator fails to work if molecule is given as minor tautomeric form.








By the way the molecule you show is the major tautomeric form?





Jozsi

Hi Jozsi,





indeed "apkb1 = 10.8 (atom 6)" should read "bpka1 = 10.8 (atom 6)".





I think it is the major tautomeric form.

Dear kai_h











Improved version of pKa calculator will be released in a couple of days.


Guanine's pKa was calculated with this improved version. See the attached figure.





Jozsi