[gmx-users] fep calculation of L-Asn - D-Asn transition

[Berk Hess]

>From: Alexandra Patriksson <alexandra at xray.bmc.uu.se>
>Reply-To: Discussion list for GROMACS users <gmx-users at gromacs.org>
>To: gmx-users at gromacs.org
>Subject: [gmx-users] fep calculation of L-Asn - D-Asn transition
>Date: Wed, 24 May 2006 11:11:38 +0200
>
>Dear all,
>
>I'm trying to do a fep calculation of the L to D transition of a amino acid 
>in a small peptide. We have done some changes to the dihedrals and the 
>charges in the topology file (see below), but still we get dH/dlambda 
>values that are far too large. Does anyone have an idea what could possibly 
>be wrong? Have we forgotten something that should have been taken into 
>account?
>
>Our changes in the toplogy file going from L -> 0 :
>
>[ atoms ]
>;   nr       type  resnr residue  atom   cgnr     charge       mass  typeB  
>   chargeB      massB
>     7  opls_293B      1    ASN     CA      2       0.25     13.019     
>opls_293B       0       13.019  ; qtot 0.94
>     8   opls_140      1    ASN     HA      2       0.06          0     
>opls_140        0       0       ; qtot 1
>     9   opls_136      1    ASN     CB      3      -0.12     14.027     
>opls_136        0       14.027  ; qtot 0.88
>    10   opls_140      1    ASN    HB1      3       0.06          0     
>opls_140        0       0       ; qtot 0.94
>    11   opls_140      1    ASN    HB2      3       0.06          0     
>opls_140        0       0       ; qtot 1
>
>[ dihedrals ]
>;  ai    aj    ak    al funct            c0            c1            c2     
>        c3
>    7    17     3     8     2            35             0             0     
>         200
>    7    17     3     9     2            -35            0             0     
>         200
>
>
>Our changes in the toplogy file going from 0 -> D :
>
>[ atoms ]
>;   nr       type  resnr residue  atom   cgnr     charge       mass  typeB  
>   chargeB      massB
>     7  opls_293B      1    ASN     CA      2          0     13.019     
>opls_293B       0.25     13.019   ; qtot 0.94
>     8   opls_140      1    ASN     HA      2          0          0     
>opls_140        0.06          0   ; qtot 1
>     9   opls_136      1    ASN     CB      3          0     14.027     
>opls_136       -0.12     14.027   ; qtot 0.88
>    10   opls_140      1    ASN    HB1      3          0          0     
>opls_140        0.06          0   ; qtot 0.94
>    11   opls_140      1    ASN    HB2      3          0          0     
>opls_140        0.06          0   ; qtot 1
>
>[ dihedrals ]
>;  ai    aj    ak    al funct            c0            c1            c2     
>        c3
>    7    17     3     8     2             0             200           -35   
>         0
>    7    17     3     9     2             0             200           35    
>         0

I don't understand why you turned off the charges.
Charge interactions are excluded anyhow for 1-3 interactions.
The problem might be that just turning of the impropers
does not have much effect, as the transition is blocked
by the angle potentials.
I would guess the D-L conversion would happen in one jump
when the improper potential is high enough to overcome
the barrier caused by the angle potentials. This would result
in a large hysteresis.
Turning of the angles (instead of the charges) might help.

Berk.