Can you describe your system? The snapshot shows four boxes, each with one protein --- is this your simulation system?

Did you align the snapshots to the reference (initial) structure prior to RMSD calculation?

Why do you think RMSD of 2 or 3 Angstroms is high?

Are there any conformational changes of active site residues that explain the value of RMSD?

---

If you want to know the effect of the ligand (or its charges) on the protein structure do the following two things:

1. MD simulation of the protein without the ligand.

2. MD simulation of the protein with the ligand having zero charge on each atom.

Hi Martin Klvana,

Thank you for your answer.

I forgot to mention in the question , that i have shown 2x2 extented box in the picture.

I have aligned the simulation structure with intial ( reference structure ) properly. still RMSd is high.

anyway, let me follow your suggestions to know whether charges of ligand are effecting or not

Thanks

anji babu

Hi Anji,

It seems that you actually have three independent questions here, and I will answer in three different points.

1) Charges for protein/ligand interactions.

     Protein ligand MD simulation are quite tricky because the interactions that will keep your drug in the binding pocket are frequently weak and can be broken along your simulation. This can either be an artefact (error in your settings) or the actual behavior that you should expect, depending on the binding affinity of your compound towards your receptor.

     From my experience, the atomic charges of your ligand will play a crucial role here. It is thus important to derive them properly. To my opinion, you should not wonder if your charges are wrong or right, but rather think about consistent or inconsistent with your choice of force field for the protein. You mentioned that you used Amber force field for the protein, you should thus refer to the original paper and follow the exact protocol herein describe (e.g., RESP calculation from B3LYP/cc-pVTZ(PCM)//HF/6-31G* electrostatic potential if you chose amber03). The level of theory chosen by the authors can always be discussed, but the most important thing is that it must be consistent.

     So I would suggest you to have a careful look at that before re-running your simulation. Notice that having a good set of charges does not necessarily mean that your simulation will be a success. Getting the charges as derived in the original force field using the RESP methodology requires running some QM calculations. If you are not familiar with those, or if you do not have access to such programs, I suggest you to use the RED server. It has been specifically developed for this purpose and it is completely free. You will need to register in order to use the Gaussian program and then chose the method that fits your problem (refer to the protein force field that you chose).

http://upjv.q4md-forcefieldtools.org/REDServer-Development/

2) High RMSD

    As Martin pointed out, 2-3 A RMSD is not particularly high, although it depends on your system. You said that this is the RMSD of the active site, which you compared to a backbone RMSD. I believe that you computed the RMSD of the active site including the residues side chains and the ligand. If so, the fluctuations of all that are, for me, compatible with such RMSD values. But once again, it depends on your system. You would probably find more relevant to identify some characteristic ligand/protein and protein/protein distances, angles and/or dihedrals that you could follow along your simulation. This would tell you clearly if your active site remains in the conformation that you expect.

      Now once again, protein/ligand MD simulations are tricky. The weak interactions that maintain your ligand in the active site can be broken during the heat up and equilibration process. You might need to try several setup and gradual heatup before you end up with anything relevant. You will find some help in the literature. 

3) The holes in the solvent box.

    This is clearly a problem. It means that you have vacuum bulbes in your box and that is not realistic. The density of your simulation box is probably way too low compared to what it should be. You should compute it to be sure. If you prepared your system with the leap program of AmberTools, it is possible that the solvation leads to such a low density. If so, one way to fix this issue can be to first minimize your simulation box and readjust the box size manually to get a proper density. A value between 0.95 and 1.05 seems to me like a good guess to start your simulation, as this will be adapted during the NPT run. If your heatup protocol is gradual and slow enough, this should not be a problem anymore.

Getting the right setup can be a long path of trials and errors, but you will eventually get something out of it.

I hope this helps. If not or if you need more information, you will find the Amber mailing list very useful. I would suggest you to register. 

Welcome to the MD community and good luck!

Antoine

Hello Anji BABU Kapakayal,

Before I give a suggestion, I would like to ask what specifically is the solvent in your system?

As for the "holes" in your system box, I think this is greatly exaggerated because of how the boxes were extended into a 2x2 box. Try extending the simulation time to 10 ns or more. This should equilibrate the system more and allow the RMSD values to "relax". If the problem still persists, then most probably there is a problem with setting up of the systems.

Cheers!

Adam

Thank you very much Stéphane Abel , Antoine Marion, Adam Jo Elatico for giving valuable suggestions.

I wanted to ask one more doubt here, Is large size of box will effect the system during simulation ??

because ,  i am using a larger box for my simulations. ie 90.0  89.0 90.0 A  aprox.

Thanks

anji babu

Hello Anji,

The large box would make the system more difficult to equilibrate as there are more interactions to consider. It would also raise a question of how well the program samples your system. A larger system would need more sampling, thus longer time to equilibrate. And yes, a 90 x 90 x 90 A box is considerably large. Try adding the least amount of solvent  enough to cover your solute (in this case the protein).

Hope that helped you.

Cheers!

Adam

The size of the box should be chosen depending on the size of the protein and cutoff for non-bonded interactions. If you use cutoff of 12 Angstroms and the protein does not undergo big conformational changes then the wall (the water-filled space between the protein and the box boundary) in your system is unnecessarily too thick, judging from your snapshot. It can be reduced --- just a guess --- by about 10 Angstroms. If you make the box smaller, the system will be much smaller and MD simulations will be much faster.