1.审稿人了解簇模型但并不精通，似乎对簇模型有敌意。他提到我们的处理方法不对，必须先使用MD进行模拟，再进行建模。但是我查看了Fahmi Himo的或类似工作,似乎并没有MD这一过程，请问在有蛋白质单晶结构的情况下，MD的模拟是必须的吗？另外，为什么审稿人会说簇模型对熵的计算不可靠？ It is surprising to observe that the authors rely solely on docking and cluster modelling without conducting simple MD simulations to confirm the stability and validity of the generated complexes. The authors crudely utilize the crystal structure, add hydrogens, and assume the crystal represents the complex well despite being originally obtained with a different ligand. Moreover, although cost-effective, docking is unreliable for predicting accurate binding poses due to its approximated scoring function. The current calculations appear rudimentary, and the agreement with experimental results is fortuitous. Cluster models are also unreliable for entropy calculation.
2.我是做实验出身，半路出家做了一些计算，并不是很专业，很多计算知识都是来自于咱们论坛。虽然看了很多论坛的文章，但是很多计算化学基本的知识还不是很清楚，希望得到大家的指导。例如：用TDDFT做电荷转移激发的计算有什么不足吗？双自由基体系必须用双杂化泛函进行计算吗？eps的选择依据是什么，必须在8-12吗？我们参考了Fahmi Himo的工作，eps是4.0。 Subsequently, the authors utilized time-dependent density functional theory (TDDFT) to explore electron transfer processes. Could they justify the methodology employed and provide benchmarks for its validity? Biradical systems are known to be challenging, and it is unclear if the methodology used is the right one for it (um062x-D3/def2tzvpp (SMD, eps= 4.0) // ub3lyp-D3BJ/ def2svp (gas)). I would expect that a double hybrid DFT would be needed for such a system. Additionally, could the authors confirm the absence of spin contamination and justify the use of such low dielectric? Often, e=8-12 is used for protein pockets.
3.簇模型计算已经明显显示tyrosine会有关键作用，我们通过实验进行了验证。但是审稿人却还是要求我们尽量用MD的方式来验证解决。是否可以用簇模型计算突变体对选择性的影响而不需要用MD？或者有其他更好的办法吗？ Finally, the authors mutated key tyrosine residues to phenylalanine to demonstrate the importance of hydrogen bonding in binding and catalysis. However, it remains unclear how these mutations affect the binding of the complex. The authors assume it remains the same, but this should be confirmed via MD simulations before proceeding with expensive DFT analyses. Such simulations can now be efficiently performed in a few days (100 ns a day). Analysis of relevant hydrogen bond interactions alongside MD simulations could also help. Such analyses are likely to be more robust and cost-effective than DFT.