ClassicalGSG: Prediction of logP Using Classical Molecular Force Fields and Geometric Scattering for Graphs

19 November 2020, Version 1
This content is a preprint and has not undergone peer review at the time of posting.


This work examines methods for predicting the partition coecient (log P) for a dataset of small molecules. Here, we use atomic attributes such as radius and partial charge, which are typically used as forcefield parameters in classical molecular dynamics simulations. These atomic features are transformed into index-invariant molecular features using a recently developed method called Geometric Scattering for Graphs (GSG). We call this approach "ClassicalGSG" and examine its performance under a broad range of conditions and hyperparameters. We train a ClassicalGSG log P predictor with neural networks using 10722 molecules from the ChEMBL21 dataset and apply it to predict the log P values from four independent test sets. The ClassicalGSG method's performance is compared to a baseline model that employs graph convolutional neural networks (GCNNs). Our results show that the best prediction accuracies are obtained using atomic attributes generated with the CHARMM generalized Force Field (CGenFF) and 2D molecular structures.


log P
parittion coefficient
machine learning
graph convolutional neural network
geometric scattering for graphs

Supplementary materials

logP SI


Comments are not moderated before they are posted, but they can be removed by the site moderators if they are found to be in contravention of our Commenting Policy [opens in a new tab] - please read this policy before you post. Comments should be used for scholarly discussion of the content in question. You can find more information about how to use the commenting feature here [opens in a new tab] .
This site is protected by reCAPTCHA and the Google Privacy Policy [opens in a new tab] and Terms of Service [opens in a new tab] apply.