colvar+ABF

LAMMPS LAMMPS Mailing List Mirror
1

Hi LAMMPS USERS,

I am running ABF calculations using colvars package for studying the dimerization free energy profile in bulk solution.

Initially, two molecules are separated by distance of 3 A centered in box of size 55A.

I have defined reaction coordinate to evolve from 2A to 15 A.

At some point two molecules moves to edge of the box and their comes the actual problem. (RC is PBC & RC is suddenly increasing to around 25A)

I can fix the first molecule center of mass to overcome this problem . will that wont effect FE profile?

orelse is there any better solution to overcome this PBC issue.

Thanks a lot in advance

Karteek Kumar

2

Hi LAMMPS USERS,

I am running ABF calculations using colvars package for studying the
dimerization free energy profile in bulk solution.

Initially, two molecules are separated by distance of 3 A centered in box of
size 55A.

I have defined reaction coordinate to evolve from 2A to 15 A.

At some point two molecules moves to edge of the box and their comes the
actual problem. (RC is PBC & RC is suddenly increasing to around 25A)

I can fix the first molecule center of mass to overcome this problem . will
that wont effect FE profile?

it would be better to (very weakly) restrain the center of mass of the
two molecules combined.

orelse is there any better solution to overcome this PBC issue.

the cleaner option would probably to implement a a "fix_modify unwrap
yes" option,
so that you get a choice over applying PBC or not.

axel.

3

Thanks Axel sir,

I will try with these options.

I guess its dump_modify unwrap yes ( sorry if I am wrong)

One more small query,

for example, I have to find the dimerization free energy of two methane in aqueous solution.

for which reaction coordinate is defined between two carbons atoms.

fix fix_id group_id colvars colvars_config.file

what should be group_id ( only methane molecules ) or it should be ‘all’

Thanks

Karteek Kumar

4

Thanks Axel sir,

I will try with these options.

I guess its dump_modify unwrap yes ( sorry if I am wrong)

that is pointless. what should that achieve? there *is* no fix_modify
option for fix colvars *yet*. it needs to be implemented, i.e.
*programmed*. you can put in your input script whatever you want, that
is not going to change the sources.

One more small query,

for example, I have to find the dimerization free energy of two methane in
aqueous solution.

for which reaction coordinate is defined between two carbons atoms.

fix fix_id group_id colvars colvars_config.file

what should be group_id ( only methane molecules ) or it should be 'all'

please *always* RT?M:

The group-ID entry is ignored. The collective variable module will
always apply to the entire system, i.e. use the group all.

the communication will be restricted to the minimum number of atoms as
defined in the colvars input file *inside* of fix colvars
automatically. no benefit from using any additional restriction on the
fix.

axel.

5

i have two more questions myself:

- why do you use ABF on a collective variable, that would be *much*
easier and more efficiently explored using umbrella sampling?

- i have implemented an option to unwrap coordinates for fix colvars
in my local branch of LAMMPS. would you mind sharing your input deck
for testing?

thanks,
   axel.