Table 4.

Different encodings from deep learning models (see Table 1 for details)

Encoding	Description	Summary	Used in	Used along with
ProtVec	amino acid sequences are encoded as a distributed representation of k-mers	Density: + Information: +	[124]
Voxel	structures of proteins are encoded as voxels	Density: o Information: +	[125, 126]
Matrix	mimicks images by regarding the respective entries of PSSMs as pixel densities	Density: o Information: +	[127, 129, 130]	PSSM
Autoencoder	extracts representative characteristics in order to reproduce the input as good as possible	Density: + Information: o	[131]