How to use MolGpKa?

1. On the website, please input a molecule (SMILES format) and then click the predict button. You can input the SMILES string directly, or you can draw a molecule through the JME molecule editor.

2. The server extracts all ionization centers on the input molecule and predicts pKa for each ionizable group by graph-convolutional neural network. The result page shows information of all ionization centers and predicted pka.

Example for pka prediction with MolGpKa API

  • MolGpka_api_example

  • To prevent overloading, please do not send over 10000 requests per day or 500 requests per hour.


    If you have any question about MolGpKa, please contact:

    Changge Ji