Skip to main content
Fig. 5 | BMC Bioinformatics

Fig. 5

From: Performance improvement for a 2D convolutional neural network by using SSC encoding on protein–protein interaction tasks

Fig. 5

The concept of SSC encoding format. a SSC encoding. Three relevant and independent matrices constitute a composite cell. Calculations of features not only possess more combinations from larger numerical ranges but are also promoted by two additional data matrices. b Details of 2D protein data conversion. According to the positions of the amino acids in the sequence, each amino acid is assigned a position number. Three channels are filled with each data point, and empty positions are addressed using zero-padding operations. The total length of 3600 × 3 can be reshaped into a 60 × 60 × 3 matrix as the input of the model

Back to article page