[gmx-users] [Fwd: ffoplsaabon.itp Help]

[David van der Spoel]

-------- Forwarded Message --------
From: Pawan Babel <pwnbabel at chem.iitb.ac.in>
To: spoel at xray.bmc.uu.se
Subject: ffoplsaabon.itp Help
Date: Wed, 27 Oct 2004 18:24:23 +0530 (IST)
Dear david

Can you please explain me why the order of the naming in ffoplsaabon.itp for
i,j ,k,l is different than usual convenction i.e. "i" is connected to "j"
and "j" is connected "k" and "k" is connected to "l" and finally (i-j-k-l)
forms a dihedral angle ?
If i connect in order of i-j-k-l for C-term phi as given below from
ffoplaabon.itp it does not form a dyhedral angle.

According to .../gromacs/top/ffoplaabon.itp

 [ dihedraltypes ]
;  i    j    k    l   func     coefficients
; OPLS Fourier dihedraltypes translated to Gromacs Ryckaert-Bellemans form
; according to the formula in the Gromacs manual.
 C_3    CT_2   N      C       3    -10.35749 -29.58716  -1.16734  41.11199
  0.00000   0.00000 ; C-term phi.

 C      CT     CT_2   NT      3      3.33465  -1.55226   2.82001  -4.60240
  0.00000   0.00000 ; H2N terminal

----------------
According to me, it should be other way round either C_3 C N CT_2
(i-j-k-l) or CT_2 N C C_3 for C-terminal phi. So that it forms a proper
dihedral angle.
And similarly CT_2 C NT CT or CT NT C CT_2 for N-terminal. If i change to
the order which i have specified the error during grompp will not occur ,
i.e. Ryckaert-Belleman paramters.

Please correct me if i am wrong and also tell me how can i calculate these
fourier constants. According to E(torsional)= Summation [v0 *(1+cos(phi))
+v1 * (1- cos(2*phi)) + v2 *(1 +cos(3*phi)]
where v0,v1,v2 are fourier coefficients in kcal/mol.
 but from where we are getting v3,v4,v5 coefficients terms in
ffoplsaabon.itp.
Any guidance and help is highly appreciable.

Thanks in advance
Pawan Babel
-- 
David van der Spoel <spoel at xray.bmc.uu.se>