RRCRank: a fusion method using rank strategy for residue-residue contact prediction.

The learning-to-rank algorithm is introduced to develop a novel rank-based method for the residue-residue contact prediction of proteins, which achieves state-of-the-art performance based on the extensive assessment.