[gmx-users] Running simulation differences

[Justin Lemkul]

On 12/26/18 3:50 AM, Gonzalez Fernandez, Cristina wrote:
> Hi Justin,
>
> Tha NAMD settings for vdw interactions are:
> Rvdw=1.2 nm
> Smooth switching function between 1 and 1.2 nm
> In this case, the CHARMM FF was used.
>
> However, one of the paper authors repeated the simulation of the same lipid but using the GROMOS simulation package. In this case, he used the GROMOS96FF and obtained very close values to the NAMD simulation. He included the topology as supporting information and I took it, thus I think my mistake comes from the input to the energy function.
> In that case the settings are:
> Pairlist generated every 5th,
> Cut-off for short range pairlist=0.8
> Cut-off for long range interactions=1.4
>
> He also included the simulations settings. I include the settings related to vdw and coulomb interactions in GROMOS in case you could help me to identify what I am doing wrong:
>
> # use grid based pairlist generation to speed up
> PAIRLIST
> #	algorithm: standard(0) (gromos96 like pairlist)
> #		             grid(1) (XX grid pairlist)
> #	SIZE:	   grid cell size (or auto = 0.5 * RCUTP)
> #	TYPE:	   chargegoup(0) (chargegroup based cutoff)
> #			         atomic(1) (atom based cutoff)
> #
> #	algorithm	  NSNB	RCUTP	RCUTL	  SIZE	TYPE
> 	        1	   5	  0.8	  1.4	   0.4	   0
> END
> # Longrange reaction field correction
> NONBONDED
> # NLRELE    APPAK      RCRF     EPSRF   NSLFEXCL
>         1      0.0       1.4        61    1
> # NSHAPE   ASHAPE    NA2CLC   TOLA2   EPSLS
>         3       1.4        2   1e-10       0
> # NKX    NKY   NKZ    KCUT
>     10     10    10     100
> # NGX   NGY   NGZ  NASORD  NFDORD   NALIAS  NSPORD
>     32    32    32       3       2        3       4
> # NQEVAL   FACCUR   NRDGRD   NWRGRD   NLRLJ    SLVDNS
>    100000      1.6        0        0       0      33.3
> END
>
>
> With this information I set in GROMACS:
> ns_type                = grid        ; search neighboring grid cells
> nstlist                = 5        ; 20 fs, largely irrelevant with Verlet
> rcoulomb            = 1.4        ; short-range electrostatic cutoff (in nm)
> rvdw                = 1.4        ; short-range van der Waals cutoff (in nm)
> rlist                   = 1.41
> vdwtype                 = Cut-off
> vdw_modifier            = None
> verlet-buffer-drift     = -1
>
>
> Could you help me to find what are the inputs of the wdw energy function that are not correct?

If you can't repeat someone else's result, you should inquire with the 
corresponding author. The most immediate task is to figure out how to 
make GROMACS produce the same result from GROMOS; the CHARMM result in 
NAMD is rather irrelevant to this investigation. I'm also surprised that 
anyone would get similar energies from GROMOS and CHARMM. They are 
entirely different functional forms and topological representations. Put 
that to the side for now.

I have no idea how to read a GROMOS input file, but again that's a 
question for the corresponding author of the paper you're reading.

-Justin

-- 
==================================================

Justin A. Lemkul, Ph.D.
Assistant Professor
Office: 301 Fralin Hall
Lab: 303 Engel Hall

Virginia Tech Department of Biochemistry
340 West Campus Dr.
Blacksburg, VA 24061

jalemkul at vt.edu | (540) 231-3129
http://www.thelemkullab.com

==================================================