发展参考势方法框架下的自由能计算方法辅助人工酶设计

现阶段，理论和计算化学手段在酶的定向进化和从头设计中所能扮演的角色非常有限。其主要的障碍在于计算机模拟研究酶反应机理需要大量的结构采样，因此非常耗时。参考势方法可以在保持从头算量子化学方法精度的前提下显著降低计算量。本项目旨在进一步发展参考势方法，使其能够解决酶设计过程中的一些具体的问题，包括残基突变导致的酶催化能力改变的快速预测、力匹配和机器学习优化参考势的哈密顿提升参考势方法的收敛速度、反应活性区域随反应进行发生动态变化、酶反应中的核量子效应等，并最终将参考势方法用于Kemp eliminase和Diels-Alderase的研究，为这两类酶的进化设计提供指导。

Currently, theoretical and computational chemistry can only play a very limited role in directed enzyme evolution and de novo design of new enzymes. The main hindrance is the formidable computational expense in large-scale samplings of configurations in phase space in order to confidently unveil the reaction mechanism. The reference-potential method can ameliorate this difficulty by significantly reduce the cost while maintaining the accuracy on ab initio levels. In this project, we aim at further improving this method and making it more suitable for solving practical issues in enzymatic design, including quick prediction of the alternation in catalytic efficiency caused by mutations, accelerating convergence by optimizing reference potential using force matching and machine learning, adaptive active region, and nuclear quantum tunneling effect in enzymatic reactions. Finally, we will apply the reference-potential method to the study of Kemp eliminase and Diels-Alderase, and provide valuable guidance for the artificial evolution of these two enzymes.