With access to large datasets, deep neural networks through representation learning have been able to identify patterns from raw data, achieving human-level accuracy in image and speech recognition tasks. However, in chemistry, availability of large standardized and labelled datasets is scarce, and with a multitude of chemical properties of interest, chemical data is inherently small and fragmented. In this work, we explore transfer learning techniques in conjunction with the existing Chemception CNN (convolutional neural network) model, to create a transferable and generalizable deep neural network for small-molecule property prediction. Our latest model, ChemNet learns in a semi-supervised manner from inexpensive labels computed from the ChEMBL database. We then fine-tune ChemNet to the Tox21, HIV and FreeSolv dataset for toxicity, activity and free energy of solvation prediction, which are 3 separate chemical tasks that ChemNet was not originally trained on. Our results indicate that with no explicit feature engineering, ChemNet exceeds the performance of existing Chemception CNN models and contemporary MLP (multi-layer perceptron) models that trains on molecular fingerprints. Furthermore, our fine-tuning experiments suggest that the lower layers of ChemNet have learned “universal” chemical representations that are generalizable to the prediction of novel and unseen small-molecule properties. Therefore, ChemNet can be used as a universal plug-and-play model, which will accelerate the deployment of deep neural networks for the prediction of novel chemical properties.

ChemNet: A transferable and generalizable deep neural network for small-molecule property prediction

255th National Meeting (2018)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.