I am trying to calculate the PMF of NaCl dissociation by doing Umbrella Sampling and further for removing bias using WHAM, Umbrella Integration (This is just for testing the codes). I checked the overlaps of umbrellas at various positions and they are in good overlap. When i tried to use WHAM, i found that the PMF generated is not in correspondence with the reported literature. Although i am getting the CIP, SSIP positions perfect but the plot shows ions like to stay in SSIP region rather than CIP which is in-correct. Further when i tried to use umbrella integration then it worked. Now, i am confused why WHAM is not working properly.
Could you clarify what WHAM code are you using? and your input?
Do you use energy or just the frequency?The two should be consist. Using frequency only should be more easy and you have few chance to get into trouble.
If you are working in multiple temperatures, to try for a single temperature first.
The tricky part could be the [1/2] in restraint spring constant. According to http://membrane.urmc.rochester.edu/sites/default/files/wham/doc.html 8.1.2.
From my own experiences, Alan Grossfield's WHAM code (http://membrane.urmc.rochester.edu/content/wham/) works quiet well for 1D PMF. It's quiet robust for same temperature and frequency only. It's easy to use, too.
Hey Qin thanks for your reply.
Well i am using lampps for simulations and by using colvar package sampling was done. Once i am done with simulations. It generates out.colvar.traj file which contain step position forces fixed-position. Using this file metadata file was generated which consists of two columns (column 1: Step-time Column 2: Instantaneous position). After which using simple WHAM (grossfield's lab) i tried to generate PMF. Things looks okay to me as far as sampling and all is considered. Do you have any idea ?
I don't find anything wrong with WHAM. I assumed your simualtion was at same temperature.
I have no idea regarding NACL association. But in the following paper Fig 2(a), ion could like to stay in SSIP for some forcefields, such as QM/EFP. Only AMBER strongly prefered CIP. What's your forcefiled? Without known it, it's hard to say the result from WHAM is totally wrong.
https://www.researchgate.net/publication/233899914_Interionic_Hydration_Structures_of_NaCl_in_Aqueous_Solution_A_Combined_Study_of_Quantum_Mechanical_Cluster_Calculations_and_QMEFP-MD_Simulations
I'm wondering whether umbrella integration or WHAM or the simulations didn't converge yet. How good is your PMF comparing with those in Fig 2(a)?
Article Interionic Hydration Structures of NaCl in Aqueous Solution:...
Hi Qin i have gone through this paper and there is another one on ab-initio dynamics for NaCl dissociation. If we talk about umbrella integration then its giving desirable results if we compare it with fig2 or other references, but WHAM shows a very small barrier of 1kcal/mol if we move towards SSIP (from CIP) and it looks like SSIP its 1.25kcal/mol stable than CIP. TIP3P water model is used in these calculations.
Sounds WHAM works as expected, but the sampling has problem.
Could your constraint be not properly, for example too strong, so that you pathway in WHAM is biased or not sufficiently sampled? Good overlap in reaction coordinate couldn't guarantee good overlap in conformation space. If sampling near the barrier is not enough, there's no way that the relative energy between CIP and SSIP would be correct.
I have no experience with umbrella integration. From what I read, the most advantage of umbrella integration is providing a better pathway automatically.