什么是constrained MD

今天看到有人在谷歌群组中自问自答，陈述了一些关于constrained MD的问题：

https://groups.google.com/forum/#!searchin/cp2k/RI|sort:relevance/cp2k/yWqahb93_38/EM9gVaNdCgAJ

以下是他回答的一些copy：

Just as these papers refer to, a constrained MD simulation is to simulate a rare event. 

Suppose for a reaction, A + B -> C, and it takes hours to happen in reality, and the reason is, A irregularly hit B, and only 1/10000 times, that is, 10000 times, A hit B once. 

Now, the constrained MD does this, it constrained the distance A-B, say, 2.5 Angstrom, and instead of 10000 times hit once, since it is constrained, in 10000 times, it hit 10000 times,  and you get 10000 times more opportunity to make the reaction to happen,  by the way, a reaction doesnot mean a "real reaction", it can mean a inter-molecular complex.

constrained MD simulations belong to potential of mean force (free energy) calculation, there are many very good reviews for free energy calculation, for example, 

"FREE ENERGY VIA MOLECULAR SIMULATION: Applications to Chemical and Biomolecular Systems D. L. Beveridge and F. M. DiCapua", Annu. Rev. Biophys. Biophys. Chem. 1989. 18:431-92, and many others. 

And Dann Frenkel and Berend Smit's book is also helpful for understand details of MD/MC simulations. 

I have lots of fun in reading and understanding, even though I am a very slow learner.

personally thinking, the constrained md is kind of biasing to the reaction coordinates, so I personally prefer running a 1000 ps unconstrained MD, to running the same length of constrained md. For some reactions with high free energy barrier, unconstrained MD cannot simulate the reaction. In this case, maybe I can try other methods.

however, at this moment, I dont think I can find a better way to simulate high free energy barrier for chemical reactions. Ab-initio constrained MD seems to be the only choice, with enough sampling and good guess for the reaction coordinate, the trajectory made up ab-initio constrained MD will be reasonable.

I think in order to use constrained MD, we need to locate the reaction coordinate, which is not an easy task. I think lots of people have been working on such problems, how to map the free energy surface, not only in ab-initio MD calculations, but also in classical MD calculations, how to locate the reactants, saddle points (transition states), and products. In quantum mechanics calculation, we use potential energy surface, in MD, we use free energy surface. 

I am still trying to learn all these, but I think in order to understand our problems, we need to have a better understanding of our energy surfaces, which is a very complicated problem.

I am not talking about using classical MD to do reactions, I am just saying that in classical MD, there are also free energy mapping, which includes local minima, maxima, and saddle points.