Skip to main content

A library to estimate bond dissociation energies (BDEs) of organic molecules

Project description


PyPI version Build Status

A machine-Learning derived, Fast, Accurate Bond dissociation Enthalpy Tool (ALFABET)

This library contains the trained graph neural network model for the prediction of homolytic bond dissociation energies (BDEs) of organic molecules with C, H, N, and O atoms. This package offers a command-line interface to the web-based model predictions at

The basic interface works as follows, where predict expects a list of SMILES strings of the target molecules

>>> from alfabet import model
>>> model.predict(['CC', 'NCCO'])
  molecule  bond_index bond_type fragment1 fragment2  ...    bde_pred  is_valid
0       CC           0       C-C     [CH3]     [CH3]  ...   90.278282      True
1       CC           1       C-H       [H]    [CH2]C  ...   99.346184      True
2     NCCO           0       C-N   [CH2]CO     [NH2]  ...   89.988495      True
3     NCCO           1       C-C    [CH2]O    [CH2]N  ...   82.122429      True
4     NCCO           2       C-O   [CH2]CN      [OH]  ...   98.250961      True
5     NCCO           3       H-N       [H]   [NH]CCO  ...   99.134750      True
6     NCCO           5       C-H       [H]   N[CH]CO  ...   92.216087      True
7     NCCO           7       C-H       [H]   NC[CH]O  ...   92.562988      True
8     NCCO           9       H-O       [H]    NCC[O]  ...  105.120598      True

The model breaks all single, non-cyclic bonds in the input molecules and calculates their bond dissociation energies. Typical prediction errors are less than 1 kcal/mol. The model is based on Keras and Tensorflow (1.x), and makes heavy use of the neural fingerprint library.

For additional details, see the (upcoming) publication:

  • St. John, P.C., Guan, Y., Kim, Y., Kim., S., and Paton, R.S., Prediction of homolytic bond dissociation enthalpies for organic molecules at near chemical accuracy with sub-second computational cost


Installation with conda is recommended, as rdkit can otherwise be difficult to install

$ conda create -n alfabet -c conda-forge python=3.7 rdkit
$ source activate alfabet
$ pip install alfabet

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for alfabet, version 0.0.5
Filename, size File type Python version Upload date Hashes
Filename, size alfabet-0.0.5-py3-none-any.whl (11.8 MB) File type Wheel Python version py3 Upload date Hashes View
Filename, size alfabet-0.0.5.tar.gz (11.8 MB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page