Li Linqing, Xu Weiwei, Lü Qiang
School of Computer Science and Technology, Soochow University, P.O.Box 158, 1 Shizi Street, Suzhou, Jiangsu, 215006, China.
Wujiang Rural Commercial Bank, Suzhou, Jiangsu, 215006, China.
J Mol Model. 2015 Nov;21(11):294. doi: 10.1007/s00894-015-2834-7. Epub 2015 Oct 30.
Computational protein-ligand docking is of great importance in drug discovery and design. Conformational changes greatly affect the results of protein-ligand docking, especially when water molecules take part in mediating protein ligand interactions or when large conformational changes are observed in the receptor backbone interface. We have developed an improved protocol, SWRosettaLigand, based on the RosettaLigand protocol. This approach incorporates the flexibility of interfacial water molecules and modeling of the interface of the receptor into the original RosettaLigand. In a coarse sampling step, SWRosettaLigand pre-optimizes the initial position of the water molecules, docks the ligand to the receptor with explicit water molecules, and minimizes the predicted structure with water molecules. The receptor backbone interface is treated as a loop and perturbed and refined by kinematic closure, or cyclic coordinate descent algorithm, with the presence of the ligand. In two cross-docking test sets, it was identified that for 8 out of 14, and 16 out of 22, test instances, the top-ranked structures by SWRosettaLigand achieved better accuracy than other protocols.
计算蛋白-配体对接在药物发现与设计中具有重要意义。构象变化极大地影响蛋白-配体对接的结果，特别是当水分子参与介导蛋白-配体相互作用时，或者当在受体主链界面观察到较大的构象变化时。我们基于RosettaLigand协议开发了一种改进的方案SWRosettaLigand。该方法将界面水分子的灵活性和受体界面的建模纳入原始的RosettaLigand中。在一个粗略的采样步骤中，SWRosettaLigand预先优化水分子的初始位置，将配体与明确的水分子对接至受体，并用水分子最小化预测结构。受体主链界面被视为一个环，并在配体存在的情况下通过运动学闭合或循环坐标下降算法进行扰动和优化。在两个交叉对接测试集中，发现在14个测试实例中的8个，以及22个测试实例中的16个中，SWRosettaLigand排名靠前的结构比其他协议具有更高的准确性。
