Re: GaMD in NAMD

From: Miao, Yinglong (miao_at_ku.edu)
Date: Thu Jun 14 2018 - 08:46:45 CDT

Hi Dhiraj,

Yes, the parameters Alex listed come with the right names (they were changed once during the implementation).

Also you may want to set “accelMDdihe on” for dual-boost GaMD simulation, which gives higher acceleration. Template input can be:

# job 1: conventional MD and GaMD equilibration
# GaMD Settings
accelMD on
accelMDdual on
accelMDdihe on
accelMDG on
accelMDGiE 1
accelMDGRestart off
accelMDGcMDSteps 1000000
accelMDGEquiSteps 25000000
accelMDGcMDPrepSteps 200000
accelMDGEquiPrepSteps 200000
accelMDOutFreq $dcdfreq
accelMDGsigma0P 6.0
accelMDGsigma0D 6.0
accelMDGRestartFile ${molname}-out.restart.gamd

run 26000000

# then production GaMD jobs:
accelMD on
accelMDdual on
accelMDdihe on
accelMDG on
accelMDGiE 1
accelMDGRestart on
accelMDGcMDSteps 0
accelMDGEquiSteps 0
accelMDGcMDPrepSteps 0
accelMDGEquiPrepSteps 0
accelMDOutFreq $dcdfreq
accelMDGsigma0P 6.0
accelMDGsigma0D 6.0
accelMDGRestartFile ${molname}-out.restart.gamd

run 25000000

=======

You can refer to this webpage for a complete description of the parameters: http://miao.compbio.ku.edu/GaMD/manual_namd.html and an illustration regarding each stage of the GaMD algorithm: http://miao.compbio.ku.edu/GaMD/manual.html.

For energetic reweighting, the previous aMD may have limitation on the system size. But GaMD with the default parameters should have no problem in recovering the original free energy profiles, provided that sufficient sampling has been achieved!

Hope this helps,
Yinglong


Yinglong Miao, Ph.D.
Assistant Professor
Center for Computational Biology and
Department of Molecular Biosciences
University of Kansas
http://miao.compbio.ku.edu

On Jun 14, 2018, at 7:40 AM, Dhiraj Srivastava <dhirajks_at_gmail.com<mailto:dhirajks_at_gmail.com>> wrote:

My protein is 25 kda.

Dhiraj

Sent from my iPhone

On Jun 14, 2018, at 1:05 AM, Francesco Pietra <chiendarret_at_gmail.com<mailto:chiendarret_at_gmail.com>> wrote:

How big was your system? As far as I know, recovering the real 3D energy surface is only allowed for small systems.
francesco

On Thu, Jun 14, 2018 at 6:15 AM, Alexander Adams <xadams_at_umich.edu<mailto:xadams_at_umich.edu>> wrote:
Hi Dhiraj,

I've had success with the following parameters for dual-boosted GaMD:

accelMD on
accelMDdihe off
accelMDdual on
accelMDOutFreq 10
accelMDG on
accelMDGcMDPrepSteps 25000000 # steps before measuring energetics for E and k0
accelMDGcMDSteps 50000000 # steps measuring energetics for E and k0
accelMDGEquiPrepSteps 25000000 # steps with fixed E and k0 based on cMD
accelMDGEquiSteps 100000000 # steps with E and k0 updated each step
accelMDGRestart on
accelMDGRestartFile GaMD.gamd

I found these by consulting the following paper:
Gaussian Accelerated Molecular Dynamics in NAMD<https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5743237/>

I believe this covers everything you are trying to specify if your configuration, though it's not entirely clear.

Best of luck,
Alex Adams

On Wed, Jun 13, 2018 at 11:09 PM, Dhiraj Srivastava <dhirajks_at_gmail.com<mailto:dhirajks_at_gmail.com>> wrote:
Hi Chitrak
                I am trying to use all the same parameter written on GaMD website but NAMD 2.13 is complaining about following parameters.
ERROR: accelMDGrestfile
ERROR: accelMDGntebprep
ERROR: accelMDGntcmdprep
ERROR: accelMDGnteb
ERROR: accelMDGntcmd
ERROR: accelMDGirest

it don't recognize these parameters.

Thanks
Dhiraj


On Wed, Jun 13, 2018 at 10:04 PM Chitrak Gupta <chgupta_at_mix.wvu.edu<mailto:chgupta_at_mix.wvu.edu>> wrote:
Hi Dhiraj,

I have played around a bit with the NAMD implementation of GaMD. Which parameters did you need help with?


Chitrak.

On Wed, Jun 13, 2018 at 10:54 PM, Dhiraj Srivastava <dhirajks_at_gmail.com<mailto:dhirajks_at_gmail.com>> wrote:
Hi
    does anyone tried GaMD with NAMD? is theres any protocol/configuration file available? I am following the tutorial on https://gamd.ucsd.edu/tutorial_namd.html. however some of the parameters, NAMD can not understand.

Thanks
Dhiraj




This archive was generated by hypermail 2.1.6 : Tue Dec 31 2019 - 23:19:59 CST