一种基于路径优化的非平衡自由能预测方法

提出一种基于路径优化的非平衡自由能预测方法。首先,通过建立蛋白质复合物解离的多目标优化模型找出一条能耗小、速度快的拉伸分子动力学SMD(Steered Molecular Dynamics)解离路径,然后,沿此路径计算解离自由能。与Jarzynski按平均力势PMF(Potentials of Mean Force)的外推方法相比,本文方法有较高的预测效率,数值算例也给出了与实验值比较的预测精度。通过拉伸分子动力学模拟还可以揭示小配体与蛋白质之间的解离全过程,为药物设计提供重要的结构信息。

A non-equilibrium method based on path optimization for predicting free energy

A non-equilibrium method based on path optimization for predicting free energy is developed in this paper. A multi-objective optimization model for unbinding protein-ligand complexes is constructed to obtain an optimal dissociation pathway with less dissociation time and lower energy consumption by using steered molecular dynamics. The unbinding free energy is then calculated along the optimal unbinding pathways. Comparing with Jarzynski extrapolation method based on potentials of mean force (PMF), the method has higher prediction efficiency. Comparisons of prediction accuracy with the experimental values are also given by numerical examples. The novel method can also simulate whole unbinding proceeding and give some important structural information for designing new drugs.