Hello brilliant people. I have some little questions about position restraint in NVT and NPT in the . I learned that only heavy atoms are applied the force to keep them in place during equilibration. So my questions are:
Could we only apply the force to the backbone chain heavy atoms of the proteins to achieve similar aim. My understanding is that proteins folding could be disrupted in equilibrition as they are made that way by non-covalent forces which are weaker than covalent bond. If we apply the restraint to the backbone atoms, the whole protein will be in good shape in the equilibration process.
I saw the temperature, pressure variations from start to the finish of the equilibration and my question is what is the starting temperature and pressure the system assume, or the system doesn’t take any assuption and just try to force the system to approach the set values? My understanding is that since the structure we have is the structure of the protein was meassured, and so the initial temperature should be the temperature at which the structure of the protein was meassured. E.g, most protein crystalls were obtained at different tempera18 DC( 291 K) or ambiant temperature (297 K) or 4 DC (277 K).
The restraints can be applied to whatever you want in the system, you just need a good reason for applying them. I am no expert in protein simulations, but my understanding is that - generally - one starts by restraining all the atoms of the protein, then gradually relaxes the side chains while still heavily restraining the backbone atoms, and lastly you take away the restraint on the backbone and obtain your unrestrained (hopefully well equilibrated) protein. How to restrain, what to restrain, for how long etc. depend a lot on the system, on the protein, and on your aims. Anyway, the literature is full of protein MD sims, you can also have a look at what tools like CHARMM-GUI produce as starting input restraints to get an idea of the workflow.
If you mean what the starting temperature should be, it depends on what to do you want to do again, e.g. if the protein is in the human body and you are interested in its properties at physiological state then you know that we have the typical temperature of ca 310K, so that’s what you want to simulate. If your question is what is the STARTING temperature, which is then gradually tuned to the value set by my thermostat, then no, it’s not the one coming from the experiments. Roughly speaking, in MD simulations the temperature is due to the speed of your atoms. Generally, when you start from scratch an MD sim you pick a temperature and the starting velocities of the atoms in your box are assigned randomly by sampling them from a Boltzmann distribution, which ensures a physical distribution of starting velocities which eventually will give rise to a starting instantaneous temperature that is what you want. For more info take a look at the velocity generation options of GROMACS. This may depend on the integration algorithm though. For pressure I honestly do not know, I guess it will depend on the barostat and the virial calculated with the first set of velocities and forces obtained, but here maybe someone more expert on the technical side will be way more useful than me.
Hi obZehn,
Thank you for your discussion. Yeah, the velocities determined the temperature and I agree. I am still kind of confusing at these notions. Maybe I need more readings and practices to get more understanding.
Yes, you can apply the restrains on the backbone atoms to get the ‘desired’ results. backbone mainly decide the folding and formation of secondary structure. The big question need to address is ‘aim’ is really leading to a physically meaningfull simulations? there are many paramters that influences the folding of proteins/peptide, did you addressed those before thinking about the restraining? As @obZehn said, the reason for applying should have a ‘good reason’, not just because the protein/peptide is not behaving the way i wanted to be. Yes, restraining during the equilibration phases help to get the protein in a good conformation.
the temperature of the protein in simulation should be based on your aim as @obZehn mentioned, the low tempeature in xray crystallograph is set to reduce the damage by xray as well as to avoid the fhermal vibration, which help to get the better resutls.