Augmenting Chemical Databases for Atomistic Machine Learning by Sampling Conformational Space.

Journal: Journal of chemical information and modeling
Published Date:

Abstract

Machine learning (ML) has become a standard tool for the exploration of the chemical space. Much of the performance of such models depends on the chosen database for a given task. Here, this aspect is investigated for "chemical tasks" including the prediction of hybridization, oxidation, substituent effects, and aromaticity, starting from an initial "restricted" database (iRD). Choosing molecules for augmenting this iRD, including increasing numbers of conformations generated at different temperatures, and retraining the models can improve predictions of the models on the selected "tasks". Addition of a small percentage of conformations (1%) obtained at 300 K improves the performance in almost all cases. On the other hand, and in line with previous studies, redundancy and highly deformed structures in the augmentation set compromise prediction quality. Energy and bond distributions were evaluated by means of Kullback-Leibler () and Jensen-Shannon () divergence and Wasserstein distance (). The findings of this work provide a baseline for the rational augmentation of chemical databases or the creation of synthetic databases.

Authors

  • Luis Itza Vazquez-Salazar
    Department of Chemistry, University of Basel, Basel CH-4056, Switzerland.
  • Markus Meuwly
    Department of Chemistry , University of Basel , Klingelbergstrasse 80 , CH-4056 Basel , Switzerland.

Keywords

No keywords available for this article.