AMBER: Distorsion in protein and ligand during MD

From: Francesco Pietra <>
Date: Sun, 13 Jan 2008 01:46:23 -0800 (PST)

With Amber 9, I am at a new protein-ligand system, coming from
amber_score_everything structure (GB in DOCK6.1). Conditions successful from
other cases are giving troubles here.

I am looking for general advice - as far as it can be provided from my
presentation - as I am in doubt whether to further carry out MD to hopefully
bring the house in order, or select different conditions from the
amber_score_everything stage on.

The protein is ca 8000 atoms (ff ff99SB), the ligand ca 125 C H N O atoms (ff
GAFF, param from Antechamber), the complex immersed in a POPC membrane solvated
TIP3P (combining in leap), some steric clashes manually removed, minimization
(no SHAKE) successful if carried out by steepest descent only, then conjugated
gradient applied. Heating to 300K:

  imin=0,irest=0, ntx=1,
  nstlim=25000, dt=0.002,
  ntpr=500, ntwx=500,
  ntt=3, gamma_ln=2.0,
  tempi=0.0, temp0=300.0,
  restraintmask=":77-521 | :POP.O2, P1, O3, O4, O1, C15, C11, N, C12, C13, C14"
  &wt TYPE='TEMP0', istep1=0, istep2=25000,
  value1=0.1, value2=300.0, /
  &wt TYPE='END' /

that is, with strong restraint on protein, ligand, and the polar head of POPC.

Equilibration by twice applying:

  imin=0, irest=1, ntx=5,
  nstlim=250000, dt=0.002,
  cut=10, ntb=2, ntp=1, taup=2.0,
  ntc=2, ntf=2,
  ntpr=1000, ntwx=1000,
  ntt=3, gamma_ln=2.0,
  restraintmask=":77-521 | :POP.O2, P1, O3, O4, O1, C15, C11, N, C12, C13, C14"

also successful, judging from Amber Tutorial A3 scripts, i.e., constant
density, temp, etot. Then, equilibration was continued for 500ps by removing
restraints on POPC. From ptraj, measure rmsd showed a constant trend after

Production, by three times applying:

  imin=0, irest=1, ntx=5,
  nstlim=250000, dt=0.0015,
  cut=10, ntb=2, ntp=1, taup=2.0,
  ntc=2, ntf=2,
  ntpr=1000, ntwx=1000,
  ntt=3, gamma_ln=2.0,

seems to be problematic, or I have a scarce understanding of what is going on,
either the ligand is moving away from the pocket, or what else:

I combined the three mdcrd with ptraj, either for the whole system, or by
stripping POP WAT and box information.

Let me tell about the latter. For reasons I don't understand, one molecule of
water was not stripped from mdcrd, although it is called WAT. Therefore, I used
that prmtop that was used for amber_score_everything, which provided the pdb
file for the above-told work with Amber (docking was carried out with a
molecule of crystallization water in the protein, so that this prmtop knows
about WAT).

RMSD Traj Tool in VMD about the loaded, combined and stripped mdcrd (add all,
backbone) was run for:

---"top", i.e. comparing with respect to frame 0 out of 549 frames. The plot
shows rmsd rising from ca 7, a small plateau at frames 100-110, a hill centered
at ca frame 350, then a slow decreasing to ca 25 for the last (549) frame.

--- "average", i.e., comparing to the average frame. The plot shows a steep
decrease of rmsd from ca 22 to ca 17 at frame 250, then a steep rise to rmsd 22
for to the last (549) frame.

I don't tell about a similar analysis for the unabridged mdcrd as I think it is
even less useful.

I wonder whether from this description a suggestion may be offered where to put
the hands.

Thanks a lot

francesco pietra

Never miss a thing. Make Yahoo your home page.
The AMBER Mail Reflector
To post, send mail to
To unsubscribe, send "unsubscribe amber" to
Received on Wed Jan 16 2008 - 06:07:02 PST
Custom Search