Dear Users ,
I am trying to run MD of a protein in a lipid environment to which a ligand
molecule is docked using the AutoDock program . I did not have any problem
when i ran MD of the protein in lipid environment without the ligand
molecule .
But i face problems once i dock the ligand molecule into the protein .
After I minimize the system (ligand-protein complex) and then I try to run
MD in the heating stage , the simulation runs without any problem ( but as
long as the wall clock is given , which should end in much shorter time )
and completes without any error in the .out file . The standard error file
consists of the following error . Same error occurs even in
water environment .The .rst file cannot be used again and the simulation
gets aborted.
forrtl: error (78): process killed (SIGTERM)
forrtl: error (78): process killed (SIGTERM)
Image PC Routine Line Source
pmemd 000000000043D95B Unknown Unknown Unknown
pmemd 0000000000407294 Unknown Unknown Unknown
pmemd 0000000000406CDC Unknown Unknown Unknown
libc.so.6 00002B80F7E22BC6 Unknown Unknown Unknown
pmemd 0000000000406BD9 Unknown Unknown Unknown
* i use 32*8 processors
The structure seems to be distroted when i visualize in VMD ( strangely this
is not the case always , sometimes the system is looks stable but when i try
to restart the simulation i get the same error ) .The system is stable
after minimization though .I dont find any error in .rst filles , they look
normal with the box information etc . But the .rst file is useless and
cannot be used to restart the simulation , and has only 3-4 frames . Same
error when i used the ibelly option for minimization or / using AM1-BCC
charge models of the ligands .
( I calculated Gaussian and RESP charges have been calculated for the ligand
molecule previously which i used as input for AutoDock for docking ) .
I could run the MD simulations of the ligands in solvent environment so I
think there is no problem with the Gaussian calculations of the ligand
molecules
This is how I prepare my input pdb file of the complex (protein+ligand in
lipid enviornment ) -
1) I use the ligand mol2 file ( for which the gaussian and RESP charges have
been calculated ) as input to dock the ligand to the receptor molecule using
AutoDock
2) Then I get the docked coordinates of the ligand from the output file
(.dlg file ) and extract the coordinates of the ligand in pdb format .
3) I then append the ligand file to the protein file to get the
ligand-protein complex .Clashing waters 2 Angstroms around the ligand
molecules are deleted
4) generation of .top and .crd files in tleap →
first I prepare the parameter file for the ligand molecule with the missing
parameters ( i followed *tutorial B4 -**TUTORIAL B4: Simulating a
pharmaceutical compound using antechamber and the Generalized Amber Force
Field*. )
*parmchk -i ligand.mol2 -f mol2 -o ligand.frcmod*
Run tleap
*tleap -f leaprc.ff99SB
source leaprc.gaff
MOL=loadmol2 ligand.mol2
loadamberparams ligand.frcmod
saveoff MOL ligand.lib
saveamberparm MOL ligand.prmtop ligand.inpcrd * (ligand .top and
.crd files )
*quit*
*tleap -f leaprc.ff99SB
source leaprc.gaff
source load_popc.in *(parameters for the ligands , lipids are
loaded ) *
loadamberparams ligand.frcmod
loadoff ligand.lib
complex =loadpdb complex.pdb * ( protein-ligand complex
pdb file loaded )
*set box complex vdw 1.0*
*saveamberparm complex.prmtop complex.inpcrd ** (save .top
and .crd files for the ligand-protein complex used as input)*
*savepdb complex complex.pdb*
*quit*
My final system consists of (protein + lipid + water + ligand ) . I
visualised the .top and .crd files using VMD and they seem ok.
Is there any mistake in the file preparation steps thats causing this
problem ? I am unable to figure out why the simulations
are unsuccessful even after many different attempts .
Any suggestions would be appreciated ,
Thanks and best regards
siddesh
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu May 05 2011 - 03:30:02 PDT
