Zwitterions // Quaternary nitrogens

[thomvet]

I just tried using the to_sigma.py script on a molecule that is zwitterionic (and simulated as such in my DFT calculations) and I ran into this line:

COSMOSAC/profiles/to_sigma.py

Line 407 in dd10c53

raise ValueError('N with a funny number of bonds: '+len(bonds))

This line flat out not does allow quaternary nitrogens - is this intentional, because COSMO-SAC was never intended to be used on such species?

My real case is more complicated, but imagine we are looking at L-glutamic acid (https://en.wikipedia.org/wiki/Glutamic_acid). This molecule is in fact in the database, but with the COOH and NH2 group in their unionized state: https://github.com/usnistgov/COSMOSAC/blob/dd10c53d5248dd26321cdecd9c77569bc1f4d3f2/profiles/UD/cosmo/WHUUTDBJXJRKMK-VKHMYHEASA-N.cosmo

In reality, in a wide range of pH values the substance is zwitterionic. My guess is that the sigma profiles of the zwitterion will look quite different from the un-ionized version of the molecule and I would expect a different phase equilibrium to be predicted.

What's the correct way of modelling this?

[thomvet]

Also, when that line is triggered the script just crashes, but that's a minor bug in comparison (can't concatenate an int with string).

[ianhbell]

Definitely the scripts were not set up to operate on such complex species. A pull request that thoughtfully addresses such a case would be welcomed by the community.

[thomvet]

I'll see what I can do.