Fig. 1
From: Predicting protein-ligand binding residues with deep convolutional neural networks
Method Outline. Each residue in the amino acid sequence is embedded in a feature space that consists of seven types of features, namely, position-specific score matrix (PSSM), relative solvent accessibility (RSA), secondary structures (SS), dihedral angle (DA), conservation scores (CS), residue type (RT) and position embeddings (PE). The dimension number d of the feature space is 30. The amino acid sequence is transformed into a feature map as the input for the deep convolutional neural network, which outputs the result of the protein-ligand binding residue prediction. Each cell represents a dimension of the feature map