PBC problem in selfcoordNum

Hi LAMMPS users,

I am running “Well-Tempered Metadynamics” using colvars with 17JAN2014 version.

I am interested to study the free energy for dimerization using collective variable as SelfCoorNum.

For a monomer, it has 3 self-contacts and for dimer ; 3 self contact for each monomer and 9 cross contacts = 15

Now, boundaries for the contacts will be 15 (dimer) and 6 (two monomers).

I have defined the switching function such that below 4 A, it is one and gradually goes to zero around 16 A. In order to get to the coordination number as 6, two monomers should be a distance more than 16 A.

After running the simulations for long enough, at some point coordination number is 6 but two monomers are very close to each other. I think some thing went bad with PBC, but not able to figure it out.

Please find the attached input and config. file

Thanks in advance


colvar.in (716 Bytes)

in.BTA.CG.nonane (913 Bytes)

This seems like a Q for Axel and the colvars package.


This seems like a Q for Axel and the colvars package.

colvars issues should be reported on the colvars home page: colvars.github.io
go to the bottom of the page and click on "Issues".
open a new issue with your description and attach your files.


Steve and Axel: Karthik emailed me directly, and we worked out the issue (or so it seems). It doesn’t seem to be a PBC issue. I recalculated the variable that he’s using on the trajectory that he ran, but without PBCs, obtaining very similar values.

It looks more like the chosen collective variable doesn’t necessarily model the system in a reversible manner, which is unfortunately a very well known problem ™ of variable-based enhanced sampling schemes.