pka_type fixes and new columns suggestions

[Shorku]

Hi!
Thank you for this amazing and very important project!
I have a bunch of suggestions for fixes (mostly regarding pka_type's), but since the entries don’t have unique identifiers, I’m not sure how to report them.

It would also be nice to have a separate column with a mixed/practical flag so as not to filter by comment text, and a relation column (like Standard Relation in ChEMBL) so that pka_value values are immediately floats.

Thanks!

[jonwzheng]

Hi, glad to hear you are finding the dataset useful -

1. there are unique identifiers for each molecular entry. (see unique_ID) but, yes, not any yet for specific individual measurements. That would be something nice to include in a future version. In any case, perhaps the unique_ID plus pka_value will be the most unique key, could you use those to describe the pka_type fixes?
2. I definitely agree that additional filtering flags would make the data easier to use - a future major release of this dataset will focus on these sorts of interoperability and parsing quality-of-life changes. If you have other ideas for specific flags that would help your workflow, please let me know

[Shorku]

Probably, a diff view is handier:
7dda0a0?w=1&diff=split

Also note, regarding the pka_type labels there are many possible collisions for example the same constant can be labeled either as pKaH(<(de)protonation site>) or as pKaH<seq_num>. The former could be handy if one is interested in learning particular groups properties, however in case only the pKa/pKaH were needed the entities with such site-specific labels would be just filtered out. E.g. for CN(C)c1ccc(C(O)(c2ccccc2)c2ccccc2)cc1 there were two alternative labels for the same constants: pKaH1 and pKaH(nitrogen).

The most straightforward way to resolve it is probably to introduce a couple of additional columns e.g. site and state to specify (de)protonated moieties and the electronic state of a molecule.

The site specific labels also need revision, e.g. there are simultaneously pKaH(carbonium) and pKaH(C+) labels.

Thank you!

[jonwzheng]

Thanks again for this, and yes you're right that site-specific information would be useful and that not all of the labels are consistent. This is simply because this is how it was reported in the original reference data (e.g. 'pK(OH)' rather than 'pKa1'). However, it takes extra time to go through those examples and reconcile the site information with (de)-protonation step, so we didn't do that in this version. It is worth doing in the future.

The 'site labels' were provided only very occasionally and haphazardly in the original print compilations. I am not sure that we will include site-specific information for every entry, because protonation sites are not actually resolvable on a micro-state level for many of the compounds - perhaps a better approach would be to enumerate all of the relevant tautomers, as is done in Leito's compilation. But that would be quite an extensive undertaking.

I haven't forgotten about your corrections - I've found some additional small things to fix and hope to lump those in with these fixes in the near future. thank you again!