[gmx-users] H-REMD in umbrella sampling

[Yuri Garmay]

Hi,

The method I suggest is this:
1) It is possible in gromacs to do HREMD using free energy parameters
in mdp and "-replex N" option in gmx mdrun.
2) It is possible to set harmonic potential length as parameter
changing with lambda
So, the goal is to set harmonic potential between COMs of molecules of interest.
It is possible to do using virtual sites.
First of all you need topology where both molecules described as one
molecule (if they are proteins it can be done with "-merge" option of
pdb2gmx, otherwise it should be done manually)
Then you have to add virtual particles in gro and top files
In gro lines will be like these:
    5DDD     DD  115   0.000   0.000   0.000 ; arbitrary x y z
    6DDD     DD  116   0.000   0.000   0.000
In top:
[ atomtypes ]
  DD   0.0   0.0 V   0.0   0.0
[ implicit_genborn_params ] ;only for gbsa
  DD   0.0 0   0.0 0.0 0.0
...
[ atoms ]
...
   115         DD      5    DDD     DD    115        0.0        0.0
   116         DD      6    DDD     DD    116        0.0        0.0

[ virtual_sitesn ] ; one site for molecule; 2 is for "center of mass"
and sequence of numbers is atoms for calculating COM position.
   115  2   1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
46 47 48 49 50 51 52 53 54 55 5
6 57

                              116  2   58 59 60 61 62 63 64 65 66 67
68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90
91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107
 108 109 110 111 112 113 114
[ bonds ]
;  ai    aj funct            c0            c1            c2            c3
  115   116     6          0.00          2000          2.00
2000 ; 6 means that harmonic potential is not chemical bond; 0.00 and
2.00 is length of bond for lambdas 0 and 1; 2000 is stiffness of bond

I don't know is this method suitable for you, but I performed some
calculation with it for small peptides using implicit solvent.
As I remember, there were some troubles with domain decomposition,
therefore for large molecules in explicit solvent it may be unuseful,
but I didn't try examine this deeply.

-- 
Best regards,
Yuri