[AMBER] Nudged Elastic Band with Explicit Solvent for a Larger Protein

From: James W. Snyder, Jr. <jsnyder3.stanford.edu>
Date: Wed, 20 Nov 2013 12:40:11 -0800 (PST)


I'm trying to run NEB for a protein with about 230 amino acids with an explicit solvent model. After minimizing the structures, I tried running NEB only to notice that the replicas tended to decrease in energy relative to the endpoints. I then went back and looked at my minimization files and noticed that the energy of the system didn't seem to decrease during minimization. If anything, it increased. When I set cut=9999.9, minimization seems to work fine, but I can't run NEB with a cut this large because it will take too long. This is the input file used for minimization,

COMT-WT-NS-solvn : free 500 sd/1000 tot minimization
imin=1 ,
ncyc=200 ,
maxcyc=1000 ,
ntb=0 ,
ntwx = 5,
cut=14.0 ,
irest=0 , ntx=1 ,
ntr=0 ,
jfastw = 4,

The file I used for NEB is,

Alanine NEB initial MD with small K
    imin = 0, irest = 0,
    ntc=1, ntf=1,
    ntpr=500, ntwx=500,
    ntb = 0, cut = 16.0, rgbmax=25.0,
    igb = 1, saltcon=0.2,
    nstlim = 40000, nscm=0,
    dt = 0.0005, ig=-1,
    ntt = 3, gamma_ln=1000.0,
    tempi=0.0, temp0=300.0,
    ineb = 1,skmin = 10,skmax = 10,
    jfastw = 4,
    fcap = 0.1,
&wt type='TEMP0', istep1=0,istep2=35000,
    value1=0.0, value2=300.0
&wt type='END'

Upon looking at this, I may not have been consistent with my use of cut, but I'm still concerned about the minimization step given that it seems to be increasing. Is it possible to run NEB with with explicit solvent and a cut < 9999.9? If so, how do I get a reasonable minimized structure to seed the NEB calculation with explicit solvent? I'm looking only for a local minima.


Jim Snyder

AMBER mailing list
Received on Wed Nov 20 2013 - 13:00:02 PST
Custom Search