ICM GUI Manual
PrevICM User's Guide
10.15 Set Formal Charges
Next

Available in the following product(s): ICM-Chemist | ICM-Chemist-Pro | ICM-VLS

To set formal charge:

About the Methods

Auto using ICM pKa Model The method for pKa prediction is unpublished/proprietary. It is a machine learning (PLS) model based on contributions of various molecular graph fragments to atomic pKa, trained on a carefully curated set of experimental pKa values.

NN Assisted pKa Model Equilibrium This is a new neural network model has been introduced for pKa prediction based on a chemical graph convolutional architecture with enhanced propagation through conjugated systems. The model captures long-range electronic effects, improving the treatment of resonance-stabilized functional groups and delocalized charge distributions. Training was performed using MolSoft’s in-house, diverse, manually curated pKa dataset.


Prev
Align/Color by 2D Scaffold
Home
Up
Next
Torsion Analysis