Using different topologies for different starting states

[Gabriel Monteiro da Silva]

Hi,

I am trying to start a WESTPA run with multiple starting states.

The starting states cover a conformational transition that was assigned with NMR, so I would have to build and equilibrate an MD system for each structure in the ensemble. 

The solute is exactly the same in all systems.

This leaves me with two choices:

1. Build each simulation system with the exact same number of solvent and ion atoms, so that the topologies are exactly the same. This has the downside that I would need to use the number of water and ion atoms that solvates the least compact state for every single system in order to avoid PBC artifacts. This would negatively impact performance as most of the starting states are very globular and wouldn't need so many solvent layers.

2. Have a different topology for each starting state, and include logic in runseg.sh to detect to which starting state each new segment belongs and get the correct topology.  Perhaps this is a very wrong assumption, but in my superficial understanding, it makes sense that as long as each system is properly solvated then there should be no artifacts introduced from the different number of water and ion atoms.

In the case of 2., I can think of methods to implement this, but I thought I'd ask if anyone in the community ever faced a similar challenge and if they could share how they did it before I spend a few days figuring it out.

Thanks!

[Lillian Chong]

Hi Gabriel,

I suggest that you go with choice #1. Even though you would be using more solvent molecules than needed for your more compact states, it is more rigorous to simulate in the same NPT ensemble, including the same number of particles N. Further, you need to use a sufficiently large box to accommodate the most expanded states during the conformational transition with at least 12A clearance from the edge of the box.

Best,

Lillian

--
You received this message because you are subscribed to the Google Groups "westpa-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to westpa-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/westpa-users/3e04065d-46dc-4d64-9701-05ce5d6f7c0dn%40googlegroups.com.

[Gabriel Monteiro da Silva]

That makes sense, I forgot about the potential for PBC artifacts from having too narrow a box in a system that expands, and in hindsight, it seems so obvious.

Thanks, Lillian!