If the structure is known and a PDB is available for the query structure, please use it as an input. With a known structure, the prediction will output binary prediction of nucleic acid binding sites and will also map the nucleic acid binding sites and binding features on the structure to give a complete view of nucleic acid binding probability. The features include Electrostatics potential, Conservation entropy, Solvation energy measured by accessible surface area, and elements of the protein architecture.
When Multiple chains form a protein complex, RBscore needs to predict each chain separately otherwise the conservation entropy calculate is potentially erroneous. As shown in the font-page, RBscore can well predict multiple chain cases. It is necessary to predict each protein chain and combine the results.
In the absence of a known structure, only the sequence is available. In this case, we can also predict nucleic acid binding sites with a machine learning strategy that is also of high accuracy. Besides, this method is extremely fast and one or more sequences can be predicted in one run.