Computational Prediction Of RNA-protein Complex Structures

RNA-protein interactions play an important role in many biological processes, like gene regulation, RNA splicing and protein synthesis. High-resolution complex structures can help us understand these biological processes. However, experimental determination of RNA-protein complex structures is tedious, expensive and time-consuming. Therefore, it’s important to develop theoretic methods to predict RNA-protein complex structures. Till now, there are few methods designed for the prediction of RNA-protein complex structures. We developed a novel protocol for the prediction of RNA-protein complex structures,3dRPC, including a RNA-protein docking procedure, RPDOCK and a RNA-protein scoring procedure, DECK-RP. The main contents of our work are described as follows.RNA-protein docking. We developed a FFT-based RNA-protein docking procedure, RPDOCK. We take full consideration of three features during the docking process. First, the atomic packing in RNA-protein interfaces is different from that in protein-protein interfaces. Secondly, electrostatics is important in RNA-protein interaction. Finally, the stacking between bases and aromatic side chains is also important in RNA-protein interaction.RNA-protein scoring. Based on previous work, we developed a distance-and environment-dependent, coarse-grained knowledge-base potential, DECK-RP, for RNA-protein scoring. In order to take multimode interaction and long-range interaction into account, we classed amino acids and nucleotides into different residue types based on their secondary structures. We used mol-fraction corrected reference state in DECK-RP.