Question about convergence in a metadynamics using TARGET

[Arthur Goetzee]

Dear fellow PLUMEDers,

I've been running a simulation using metadynamics to obtain a FES of two beta sheets attaching and detaching, but I am uncertain about it's convergence.

I am using TARGET in conjunction with a bunch of distances between atoms as my CV. The reference distances correspond to a native state of my structure, hence the closer TARGET gets to 0, the closer I am to my native state.

When running the metadynamics simulation using this CV, I sometimes observe a large minimum the closer I get to the native state. But other times, this minimum is not present at all and the FES is flat.

When checking for convergence I see some very odd behavior. The minimum appears at some point, but soon after the system starts sampling other values for my CV for so long that the minimum dissapears. I've included an animation of the FES through time to make this more insightful.

My question is: Is possible for this system to converge, or is my CV not suitable for this?

Many thanks for your help!

Arthur Goetzee