Fig. 1
From: DeepEGFR a graph neural network for bioactivity classification of EGFR inhibitors
Overview of the DeepEGFR model architecture and data curation workflow. It illustrates the DeepEGFR framework, showcasing the integration of SMILES-based input and molecular fingerprints for classifying compounds into three activity classes: Active, Inactive, and Intermediate. The architecture includes Graph Attention Layers (GAT Layer 1 and GAT Layer 2), followed by a global pooling layer to aggregate node features. The processed data is then fed into the Graph Neural Network (GNN) for activity classification. Additionally, the figure highlights the data curation process, including the distribution of compounds across activity classes.
