Transfer Learning Graph Representations of Molecules for pKa, 13C-NMR, and Solubility

We explore transfer learning models from a pre-trained graph convoluntional neural network representation of molecules, obtained from SchNet, 1 to predict 13 C-NMR, pKa, and logS sol- ubility. SchNet learns a graph representation of a molecule by associating each atom with an “embedding vector” and interacts the atom-embeddings with each other by leveraging graph- convolutional filters on their interatomic distances. We pre-trained SchNet on molecular energy and demonstrate that the pre-trained atomistic embeddings can then be used as a transferable representation for a wide array of properties. On the one hand, for atomic properties such as micro-pK1 and 13 C-NMR, we investigate two models, one linear and one neural net, that inputs pre-trained atom-embeddings of a particular atom (e.g. carbon) and predicts a local property (e.g. 13 C-NMR). On the other hand, for molecular properties such as solubility, a size-extensive graph model is built using the embeddings of all atoms in the molecule as input. For all cases, qualitatively correct predictions are made with relatively little training data (< 1000 training points), showcasing the ease with which pre-trained embeddings pick up on important chemical patterns. The proposed models successfully capture well-understood trends of pK1 and solu- bility. This study advances our understanding of current neural net graph representations and their capacity for transfer learning applications in chemistry.