free energy-fix smd

Dear Lammps users,
I have a question about free energy calculations using “fix smd”. I would be appreciated if you can help.

I want to calculate the free energy of adsorption of a peptide on a surface. Hence, I used “fix smd” to apply a (harmoic) force on one residue of this peptide to pull it off of the surface. Now I have two questions: (1) once I apply a force to a group of atoms of the peptide, surface starts to move as well! should I fix the surface? doesn’t it affect the forces?
(2) If we apply force on different groups of atoms (e.g. different aa in the peptide) the pmf would be different. How should I deal with these? (just if anyone who has done this before can give me an idea, I would be really appreciated)

Thanks for your help in advance

Dear Lammps users,
I have a question about free energy calculations using "fix smd". I would be
appreciated if you can help.

I want to calculate the free energy of adsorption of a peptide on a surface.
Hence, I used "fix smd" to apply a (harmoic) force on one residue of this
peptide to pull it off of the surface. Now I have two questions: (1) once I
apply a force to a group of atoms of the peptide, surface starts to move as
well! should I fix the surface? doesn't it affect the forces?

you didn't say *how* you apply fix smd. i don't see how your
surface could be moving. if you're doing a proper slab geometry,
you'd be fixing a lower layer of the slab in space in any case
to have a proper representation of a bulk system, where you
cut it off, instead of having to account for surface relaxation
and reconstruction on both sides.

(2) If we apply force on different groups of atoms (e.g. different aa in the
peptide) the pmf would be different. How should I deal with these? (just if
anyone who has done this before can give me an idea, I would be really
appreciated)

why don't you apply the steering force to the entire protein?
that would avoid any distortion from inducing structural
changes where you are pulling. also, you have to be very
careful in the choice of the pulling speed. since you have
a very simple reaction coordinate and a fairly simple transformation,
wouldn't it be easier to apply umbrella sampling and thus
have better controllable statistics and an embarrassingly
parallel project. of course you can "seed" the individual
bins for the umbrella sampling conveniently from doing
an smd run.

cheers,
    axel.

Thank you very much for your respond,
here is the input script (this is just for a single water molecule pulling off of the surface as a test)

group pull id 253 254 255 # water atoms
group no_move molecule 0 # molecule 0 is the surface
fix 1 pull smd cvel 20.0 -0.00007 tether NULL NULL 20.0 0.0
fix 2 no_move momentum 1 linear 0 0 1
fix 3 all nvt temp 298.15 298.15 100.0
thermo_style custom step etotal ke temp pe ebond eangle edihed eimp evdwl ecoul elong press pxx pyy pzz f_1[1] f_1[2] f_1[3] f_1[4] f_1[5] f_1[6] f_1[7]

As the water is being pulled, the surface is moving upward too!
I don’t like to make the surface rigid and just want to prevent it from moving upward, that is why I used the “fix momentum”. But apparently I am doing something wrong.

Thanks again for your help and advise
Arshia

Thank you very much for your respond,
here is the input script (this is just for a single water molecule pulling
off of the surface as a test)

group pull id 253 254 255 # water atoms
group no_move molecule 0 # molecule 0 is the surface
fix 1 pull smd cvel 20.0 -0.00007 tether NULL NULL
20.0 0.0
fix 2 no_move momentum 1 linear 0 0 1
fix 3 all nvt temp 298.15 298.15 100.0
thermo_style custom step etotal ke temp pe ebond eangle edihed eimp
evdwl ecoul elong press pxx pyy pzz f_1[1] f_1[2] f_1[3] f_1[4] f_1[5]
f_1[6] f_1[7]

As the water is being pulled, the surface is moving upward too!
I don't like to make the surface rigid and just want to prevent it from

you don't have to make the surface itself rigid.
it is very difficult to discuss this without knowing
what kind of surface this is and how you model it.
i am assuming that you have some kind of slab
system that represents a bulk object. in that case,
you need to have a bit more elaborate setup to
properly represent the material. it is a bit unusual
that by pulling a water molecule you can pull an
apparently much larger object as well.

moving upward, that is why I used the "fix momentum". But apparently I am

fix momentum is a bad choice. depending on how
you'll set it up it can do very unphysical things to
your simulation.

doing something wrong.

if you want to pull the water away fromt he surface,
you have to use fix smd in couple mode.

what you do right now is to pull on the water
relative to a fixed position in space. this will
indirectly pull on the surface. rather you want
to get the force between the surface and the
water, which means you have to set up the
pulling accordingly.

cheers,
    axel.

Thanks again, problem solved. I did not know that the order of fixes are important
Arshia