Graph networks for molecular design

Deep learning methods applied to chemistry can be used to accelerate the discovery of new molecules. This work introduces GraphINVENT, a platform developed for graph-based molecular design using graph neural networks (GNNs). GraphINVENT uses a tiered deep neural network architecture to probabilistically generate new molecules a single bond at a time. All models implemented in GraphINVENT can quickly learn to build molecules resembling the training set molecules without any explicit programming of chemical rules. The models have been benchmarked using the MOSES distribution-based metrics, showing how GraphINVENT models compare well with state-of-the-art generative models. This work compares six different GNN-based generative models in GraphINVENT, and shows that ultimately the gated-graph neural network performs best against the metrics considered here.

Automated summary

Deep learning methods applied to chemistry have accelerated the discovery of new molecules, with GraphINVENT as a platform using graph neural networks to design new molecules. The models in GraphINVENT can quickly learn to generate molecules resembling the training set molecules without explicit programming of chemical rules, and have been compared with state-of-the-art generative models using MOSES distribution-based metrics. The study found that the gated-graph neural network performs the best among the six different GNN-based generative models in GraphINVENT.