Verification of the rule max(LpKaH) < min(LpKa)

[pkrezel]

Hi,

Thank you for this dataset.
I wanted to see if all the molecules verify the rule: max(LpKaH) < min(LpKa)

In fact it is the case, except for this molecule:

Is it an error ?

Pascal

[jonwzheng]

hi Pascal, thank you for catching this. This does look like a mistake. The other examples you list here also seem problematic because their numerical values are nearly identical (and so they should have the same label). I would recommend leaving all of these data points out of any modeling efforts for now and we will correct them in a future release. Thank you again for spotting this.

[pkrezel]

Hi Jonathan,
Thank you for your answer.
Maybe all of these molecules are in fact zwiterions ?

I found some other problems:
for smiles="CN1C=C(O)C=C[NH2+]1.[Cl-]", we get a pKaH but it is the acid conjugate I suppose.

Pascal

[pkrezel]

Hi Jonathan,

Some ionic molecules were given without counter-ion.does-it mean that the anion is OH- and that these molecules can't have neutral state in water ? (pKa > 14)

Pascal

[pkrezel]

Hi Jonathan,

One positiv ionic molecules is given without counter-ion:

Is it an error ?
I think it is a zwitterion.

Pascal

[pkrezel]

Hi Jonathan,

Is it possible that this molecule can be weak acid:

pKa = 9.2
I think it is an error, it is pKaH=9.2

Pascal

[pkrezel]

Hi Jonathan,

What does pKR mean ?

Pascal

[jonwzheng]

Hi Pascal,

Hi Jonathan,
Thank you for your answer.
Maybe all of these molecules are in fact zwiterions ?

In this dataset, the convention we have used is that we list the SMILES of the neutral isomer even if zwitterions are present. The species you listed previously indeed are problematic because the same numerical value should not correspond to multiple acidity types; we will correct that soon.

Some ionic molecules were given without counter-ion.does-it mean that the anion is OH- and that these molecules can't have neutral state in water ? (pKa > 14)

I would check the "IUPAC name" and "nickname" or "remarks" of the corresponding entry to see if there is any additional context provided. If not, then I would assume that it is indeed the acid dissociation of an ion (or equivalently, a 2nd acid dissociation of a neutral acid).

One positiv ionic molecules is given without counter-ion:

The example you have listed appears to be just an anion that can act as a base. For instance, the pKaH1 of that species NCCOS(=O)(=O)[O-] is the pKa of NCCOS(=O)(=O)O. (However, I think you are right that the neutral form should probably exist predominantly as a zwitterion)

Is it possible that this molecule can be weak acid:

I think you are right that this is likely a base. We will re-examine this datapoint to see if it needs to be updated.

What does pKR mean ?

pKR is a specific type of dissociation constant for aromatic compounds. Rather than proton loss it represents the hydrolysis reaction, e.g. R+ + H2O <=> ROH + H+.

[jonwzheng]

Hi @pkrezel the latest version of the dataset fixed these entries as well as some similar issues for ~150 other compounds in which the acidity type was incorrect. Thank you for identifying these problematic entries!