Department of Physics Huazhong University of Science and Technology, Wuhan 430074, Hubei, China.
Sci Rep. 2013;3:1887. doi: 10.1038/srep01887.
Three-dimensional structures of RNA-protein complexes are crucial for understanding their diverse functions. However, the number of the RNA-protein complex structures solved by experiments is still limited at present. To solve this problem, some computational protocols have been proposed to predict three-dimensional RNA-protein complex structures. But the prediction accuracies of these protocols are lower. The reason may be that these protocols don't fully incorporate the features of RNA-protein interfaces. Here we propose a novel computational protocol for three-dimensional RNA-protein complex structure prediction, 3dRPC, which applies new schemes to the discreteness of molecule and charge in docking algorithm and the construction of the reference state in scoring function in order to take account of the features of RNA-protein interfaces. This protocol achieves a high accuracy comparable to the well-developed algorithms for three-dimensional structure prediction of protein-protein complexes when tested on a RNA-protein docking benchmark.
RNA-蛋白质复合物的三维结构对于理解它们的多种功能至关重要。然而，目前通过实验解决的 RNA-蛋白质复合物结构的数量仍然有限。为了解决这个问题，已经提出了一些计算方案来预测三维 RNA-蛋白质复合物结构。但是，这些方案的预测精度较低。原因可能是这些方案没有充分结合 RNA-蛋白质界面的特征。在这里，我们提出了一种新的用于三维 RNA-蛋白质复合物结构预测的计算方案 3dRPC，它将新方案应用于对接算法中的分子和电荷的离散性以及评分函数中的参考状态的构建，以考虑 RNA-蛋白质界面的特征。在 RNA-蛋白质对接基准测试中，该方案在测试蛋白质-蛋白质复合物三维结构预测的成熟算法时，达到了可与之一比的高精度。
