Skip to main content
. Author manuscript; available in PMC: 2021 Jul 9.
Published in final edited form as: Chem. 2020 Jun 16;6(7):1527–1542. doi: 10.1016/j.chempr.2020.05.014

Figure 2.

Figure 2.

The illustration of graph convolutional networks (GCN) with different representations of the caffeine molecule as input. Molecular information can be represented as atomic and bond feature tensors extracted from connectivity based 2D information, or as distance matrices obtained from 3D coordinates, or any other form of sensible chemical representations.