DeepSacon is a deep neural network learning-based method for both contact number and solvent accessibility prediction by incorporating additional structure related features and amino acid related features. It can model complex sequence-structure relationship by a deep hierarchical architecture.

The source code and data of DeepSacon are available at: download