On Wed, Jul 3, 2013 at 8:47 PM, Biao Ma <jackyma1981.gmail.com> wrote:
> Hello, amber users,
>
> Now, I am reading a paper that is "Insights into structural properties of
> denatured human prion 121-230 at melting temperature studied by replica
> exchange molecular dynamics". Base on the describe of paper, I write the
> process and the input file of amber simulation, but I cannot sure that is
> correct.****
>
> Who can help me check it?****
>
I'll offer some help and comments, but on the whole there is too much here
to offer an in-depth evaluation of your procedure in a reasonable amount of
time. In general, the best way to validate a specific scheme is to try it
on some systems, especially ones where you already know what the answer
should be.
> **
> Thank a lot.
> Biao Ma
> **
> pumed url of the paper: http://www.ncbi.nlm.nih.gov/pubmed/22339436****
>
> ** **
>
> Blow is a part of method in the paper:****
>
> =====start
> In this work, the temperature setting used for sheep PrP 125-230 such that
> MD simulations were exchanged at 320.0, 322.0,324.0, 326.0, 328.1, 330.2,
> 332.3, 334.4, 336.5, 338.6, 340.8, 343.0, 345.2, 347.4, 349.6, 351.8,
> 354.0, 356.2, 358.5, 360.8, 363.1, 365.4, 367.7, and 370.0 K was used. All
>
These temperatures appear to be evenly spaced. I would suggest using a
geometric temperature series to optimize the performance of replica
exchange. I have found this server [
http://folding.bmc.uu.se/remd/] to be
remarkably accurate.
> of the replicas were equilibrated for 20 ns without exchanging temperatures
> and then extended for 65 ns of REMD simulation. The generalized Born model
> used in this study modified the calculation of Born radii and improved the
> accuracy in the solvent polarization for macromolecules. The combinational
> use of the all-atom point-charge force-field (also known as ff03) and the
> generalized Born model led to successful folding of several proteins. The
> AMBER 11 simulation package 26 was used in both REMD simulation and data
> analysis. The melted huPrP 121-230 was computed starting from an extended
> huPrP. To generate the initial extended structure, a heating method was
> applied to a known NMR structure (PDBcode 1hjn,15Figure 1A), enabling it to
> unfold at 600 K for 40 ns of MD simulation to result in an extended
> conformation (Figure 1B) as described previously.****
>
> During this simulation, the disulfide covalent bond between residues 179
> and 214 was preserved. In total, 24 replicas with duration of 65 ns and
> with an integration time step of 2 fs were computed based on the extended
> huPrP with different random number seeds to generate the initial
> conditions. A 16 Å force-shifted non-bonded cutoff and generalized Born
> solvent models with salt concentration of 0.2 M were applied.****
>
> **=====end
> **
>
> **
>
> **
>
> Simulation Procedure:****
>
> 1. system building****
>
> 2. system minimization****
>
> 3. heating system****
>
> 4. generate the extend conformation****
>
> 5. local minimization after heating system****
>
> 6. equilibrate the every replica****
>
> 7. REMD simulation****
>
> ** **
>
> ** **
>
> *1.leap.inp for system building*
>
> ** **
>
> source pdb: 1ag2****
>
> use the the ff03 (Duan et al.) force field****
>
> ** **
>
> leap.inp
>
> source leaprc.ff03.r1****
>
> set default PBradii mbondi2****
>
> # load pdb file****
>
> 1ag2 = loadPdb input.pdb****
>
> savePdb 1ag2 1ag2.pdb****
>
> bond 1ag2.179.SG <http://1ag2.179.sg/> 1ag2.214.SG <http://1ag2.214.sg/
> >****
>
> # save 1ag2 to prmtop and inpcrd files****
>
> saveAmberParm 1ag2 1ag2.prmtop 1ag2.inpcrd****
>
> # finish****
>
> quit****
>
> ** **
>
> 2. system minimization****
>
> minimisation for heated system****
>
> &cntrl****
>
> imin=1, maxcyc=1000, ncyc=500,****
>
> igb=5, ntb=0,****
>
> cut = 16, rgbmax = 16, saltcon = 0.2,****
>
This cutoff is too small IMO. I've seen people use a cutoff of 30 Å when
employing a salt concentration like the one you have, but I would suggest
using an infinite cutoff with a GB simulation.
>
> ntpr=100****
>
> /****
>
> ~****
>
> ~****
>
> ** **
>
> *3. heat the system*
>
> heating system from 0 K to 600K during 100 ps.****
>
> &cntrl****
>
> nstlim = 50000, dt = 0.002,****
>
At 600 K, a 2 fs time step may be too short.
> ntt = 1, tautp = 1.0,****
>
I would suggest using Langevin dynamics to maintain your temperature,
since it is more theoretically rigorous than the Berendsen thermostat.
ntt = 3, gamma_ln=1.0, ig=-1,
> tempi = 0, temp0 = 600, ntc =2, ntf = 2,****
>
> ntpr =100, ntwx = 100,****
>
> ntb = 0, igb = 5, ****
>
> cut = 16, rgbmax = 16, saltcon = 0.2,****
>
> nmropt = 1,****
>
> /****
>
> &wt****
>
> type = 'TEMP0', istep1 = 0, istep2 =50000, value1 = 0, value2=600,****
>
> /****
>
> &wt type = 'END'****
>
> /****
>
> ** **
>
> *4**. **generate the full unfolded conformation*
>
> ** **
>
> To generate the initial extend structure, a heating method was used to a
> known NMR structure (PDB code:1ag2), enabling it to unfold at 600 K for 40
> ns of LD simulation to result in an extended conformation.****
>
> here, I do not sure whether I do the restart MD(set irest = 1, ntx = 5, or
> irest = 0, ntx = 1), Simulation results is difference, which setting should
> I use?
>
Since the last simulation ended at 600 K, you should just restart. The
results will actually be different between any two runs (generally, unless
you run 2 identical sets of inputs on the same model GPU).
>
> 40nsld.inp****
>
> enabling the heated NMR structure to unfold at 600 K for 40ns of LD
> simulation****
>
> &cntrl****
>
> irest = 1, ntx = 5,****
>
> nstlim = 20000000, dt = 0.002,****
>
> ntt = 3, gamma_ln = 1.0,****
>
Make sure you set an explicit value for ig (the random number seed). It
should be different for every simulation or you will begin to see
synchronization artifacts.
http://pubs.acs.org/doi/abs/10.1021/ct800573m
Setting ig=-1 will pull the random seed from the wall clock, ensuring that
they are different for each run.
>
> tempi = 600,temp0 = 600,****
>
> ntb = 0, igb = 5,****
>
> ntpr = 500, ntwx = 1000, ntwr = 2000000,****
>
> ntc = 2, ntf = 2,****
>
> cut = 16, rgbmax = 16, saltcon = 0.2,****
>
> /
>
>
> 600kmd.inp****
>
> enabling the heated NMR structure to unfold at 600 K for 40ns of LD
> simulation****
>
> &cntrl****
>
> irest = 0, ntx = 1,****
>
> nstlim = 20000000, dt = 0.002,****
>
> ntt = 3, gamma_ln = 1.0,****
>
> tempi = 600,temp0 = 600,****
>
> ntb = 0, igb = 2,****
>
> ntpr = 500, ntwx = 1000, ntwr = 2000000,****
>
> ntc = 2, ntf = 2,****
>
> cut = 16, rgbmax = 16, saltcon = 0.2,****
>
> /
>
> 40nsld.out irest = 1, ntx = 5,****
>
> 600kmd.out irest = 0, ntx = 1****
>
> [image: 40nsld.etot.png]****
>
> [image: 600kmd.etot.png]
>
>
> ** **
>
> 5. local minimization after heating system****
>
> use the mdin file which is same to step 3 above。****
>
> ** **
>
> 6. equilibrate the every replica****
>
> ** **
>
> equilibrate.mdin****
>
> ** **
>
> equilibration 20 ns, every 10ps save output. <- Is this must to do
> equilibrate for 20 ns?**
>
> equilibration****
>
> &cntrl****
>
> irest=0, ntx=1,****
>
> nstlim=10000000, dt=0.002,****
>
> irest=0, ntt=3, gamma_ln=1.0,****
>
> temp0=XXXXX, ig=RANDOM_NUMBER,****
>
You can just use ig=-1.
>
> ntc=2, ntf=2, nscm=1000,****
>
> ntb=0, igb=5,****
>
> cut = 16, rgbmax = 16, saltcon = 0.2,****
>
> ntpr=5000, ntwx=5000, ntwr=10000000,****
>
> nmropt=1,****
>
> /****
>
> &wt TYPE='END'****
>
> /****
>
> DISANG=system_chir.dat****
>
> ** **
>
> 7. REMD simulation****
>
> remd 65ns exchange every 2ps <- Here I have been confused, how often to
> exchange is more appropriate ?****
>
For T-REMD, you don't lose any efficiency by attempting exchanges more
often. I would attempt exchanges at least every 100 steps, although we've
shown that convergence gets better for even more rapid exchange attempts.
>
> ** **
>
> remd.mdin****
>
> remd 65ns exchange every 2ps****
>
> &cntrl****
>
> irest=0, ntx=1,****
>
> nstlim=1000, dt=0.002,****
>
> irest=0, ntt=3, gamma_ln=1.0,****
>
> temp0=XXXXX, ig=RANDOM_NUMBER,****
>
> ntc=2, ntf=2, nscm=1000,****
>
> ntb=0, igb=5,****
>
> cut = 16, rgbmax = 16, saltcon = 0.2,****
>
> ntpr=100, ntwx=1000, ntwr=100000,****
>
> nmropt=1,****
>
> numexchg=32500,****
>
> /
>
HTH,
Jason
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Jul 04 2013 - 07:30:02 PDT
