[gmx-users] distance_restraints

Justin Lemkul jalemkul at vt.edu
Sun Jun 9 17:15:11 CEST 2013



On 6/9/13 5:15 AM, maggin wrote:
> Hi, all
>
> I use GMX4.5.5,  GROMOS96 53a6 force field ro simulation 1dx0.pdb, I use two
> steps energy minimization (steep and cg ) in vacuum as follows:
>
> 1. pdb2gmx -f 1dx0.pdb -o xxx.gro -ignh -ter -water  spce  -ss  -p xxx.top
>
> 2. editconf -f xxx.gro -o xxx.pdb -c -d 0.9 -bt cubic
>
> 3.grompp -f em.mdp -c xxx.pdb -p xxx.top -o xxx.tpr
>
> 4.mdrun -v -s xxx.tpr -nt 2 -deffnm em_xxx
>
> 5.grompp_d -f cg.mdp -c em_xxx.gro -p xxx.top -o xxx.tpr
>
> 6.mdrun_d -v -s xxx.tpr -nt 2 -deffnm cg_xxx
>
> steep is no problem, while cg have some warnings:
>
> Step 143, Epot=-1.190781e+04, Fnorm=3.680e+01, Fmax=7.027e+02 (atom 217)
> Step 144, Epot=-1.190936e+04, Fnorm=2.620e+01, Fmax=2.688e+02 (atom 217)
> Step 145, Epot=-1.191034e+04, Fnorm=2.618e+01, Fmax=2.276e+02 (atom 277)
> Step 146, Epot=-1.191235e+04, Fnorm=3.672e+01, Fmax=4.235e+02 (atom 277)
> Step 147, Epot=-1.192020e+04, Fnorm=6.482e+01, Fmax=8.288e+02 (atom 219)
> Step 148, Epot=-1.196346e+04, Fnorm=1.315e+02, Fmax=1.971e+03 (atom 378)
>
> Step -1, time -0.002 (ps)  LINCS WARNING
> relative constraint deviation after LINCS:
> rms 0.000694, max 0.004189 (between atoms 296 and 297)
> bonds that rotated more than 30 degrees:
>   atom 1 atom 2  angle  previous, current, constraint length
>
> Step -1, time -0.002 (ps)  LINCS WARNING
> relative constraint deviation after LINCS:
> rms 0.009008, max 0.056534 (between atoms 1 and 4)
> bonds that rotated more than 30 degrees:
>   atom 1 atom 2  angle  previous, current, constraint length
>      181    182   31.3    0.1000   0.1000      0.1000
>      685    688   37.8    0.1000   0.1051      0.1000
>      262    263   41.0    0.1000   0.1000      0.1000
>        1      4   39.2    0.1000   0.1057      0.1000
> Step 149, Epot=-1.197948e+04, Fnorm=1.069e+02, Fmax=1.271e+03 (atom 218)
> Step 150, Epot=-1.198617e+04, Fnorm=6.868e+01, Fmax=9.492e+02 (atom 218)
> Step 151, Epot=-1.199312e+04, Fnorm=6.069e+01, Fmax=8.678e+02 (atom 378)
> Step 152, Epot=-1.199645e+04, Fnorm=2.884e+01, Fmax=3.323e+02 (atom 378)
> Step 153, Epot=-1.199771e+04, Fnorm=2.482e+01, Fmax=2.873e+02 (atom 378)
> Step 154, Epot=-1.199942e+04, Fnorm=3.681e+01, Fmax=6.652e+02 (atom 378)
> Step 155, Epot=-1.200439e+04, Fnorm=7.122e+01, Fmax=1.091e+03 (atom 378)
> .
> .
> .
> .
> .
> Step 208, Epot=-1.231153e+04, Fnorm=2.409e+01, Fmax=2.882e+02 (atom 262)
> Step 209, Epot=-1.231272e+04, Fnorm=2.635e+01, Fmax=3.012e+02 (atom 261)
> Step 210, Epot=-1.231523e+04, Fnorm=3.957e+01, Fmax=7.031e+02 (atom 378)
> Step 211, Epot=-1.232519e+04, Fnorm=6.187e+01, Fmax=1.273e+03 (atom 378)
>
> Step -1, time -0.002 (ps)  LINCS WARNING
> relative constraint deviation after LINCS:
> rms 0.000001, max 0.000005 (between atoms 775 and 776)
> bonds that rotated more than 30 degrees:
>   atom 1 atom 2  angle  previous, current, constraint length
>      262    263   30.3    0.1000   0.1000      0.1000
> Step 212, Epot=-1.233691e+04, Fnorm=4.926e+01, Fmax=5.858e+02 (atom 378)
> Step 213, Epot=-1.234682e+04, Fnorm=8.846e+01, Fmax=1.130e+03 (atom 261)
> Step 214, Epot=-1.235149e+04, Fnorm=8.846e+01, Fmax=1.372e+03 (atom 262)
>

What does a visual inspection of your structures (initial and final) and the 
trajectory tell you about these problematic areas?

> writing lowest energy coordinates.
>
> Back Off! I just backed up em_cg_constrain_vacuum.gro to
> ./#em_cg_constrain_vacuum.gro.1#
>
> Polak-Ribiere Conjugate Gradients converged to Fmax < 100 in 275 steps
> Potential Energy  = -1.25808627767813e+04
> Maximum force     =  9.11707505025239e+01 on atom 757
> Norm of force     =  1.65933958355887e+01
>

The outcome appears quite good overall, but one should never ignore LINCS 
warnings, so you do need to carefully evaluate what happened visually.

> I constraint h-bond in cg.mdp:
> title               =  bovin
> cpp                 =  /usr/bin/cpp
> define              =  -DFLEXIBLE
> constraints         =  h-bonds
> integrator          =  cg
> dt                  =  0.002    ; ps !
> nsteps              =  2000
> nstlist             =  10
> ns_type             =  grid
> rlist               =  1.0
> coulombtype         =  PME
> rcoulomb            =  1.0
> vdwtype             =  cut-off
> rvdw                =  1.4
> fourierspacing                =  0.12
> fourier_nx                =  0
> fourier_ny                =  0
> fourier_nz                =  0
> pme_order                =  4
> ewald_rtol                =  1e-5
> optimize_fft                =  yes
> emtol               =  100
> emstep              =  0.001
> ; Highest order in the expansion of the constraint coupling matrix
> lincs_order              = 8
>
> ; GENERATE VELOCITIES FOR STARTUP RUN
> gen-vel                  = no
> gen-temp                 = 293
> gen-seed                 = 173529
>
> in order to fix warning, I use distance_restraints as follows:
>
> genrestr  -f em_xxx.gro  -o posre.itp -disre_dist 0.1 -disre_up2 1
>
> acording to the posre.itp, revise xxx.top file,
>
> ; Include Position restraint file
> #ifdef POSRES
> #include "posre.itp"
> [ distance_restraints ]
> ; ai aj type index type’ low up1 up2 fac
> 1 4 1 0 2 0.0 0.19 1.19 1.0
> 181 182 1 0 2 0.0 0.19 1.19 1.0
> 262 263 1 0 2 0.0 0.19 1.19 1.0
> 296 297 1 0 2 0.0 0.19 1.19 1.0
> 353 354 1 0 2 0.0 0.23 1.23 1.0
> 512 516 1 1 2 0.0 0.24 1.24 1.0
> 685 688 1 0 2 0.0 0.19 1.19 1.0
> 775 776 1 0 2 0.0 0.20 1.20 1.0
> 994 995 1 2 2 0.0 0.24 1.24 1.0
> #endif
>
> But things not changed, same warning
>

The distance restrains are ineffectual - you're allowing a 1-nm region where 
there is no change in potential, so what purpose would these restraints serve? 
In any case, the LINCS warnings indicate that there is something wrong with the 
starting geometry, so trying to preserve that bad geometry is a bad idea in all 
likelihood.

-Justin

-- 
========================================

Justin A. Lemkul, Ph.D.
Research Scientist
Department of Biochemistry
Virginia Tech
Blacksburg, VA
jalemkul[at]vt.edu | (540) 231-9080
http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin

========================================



More information about the gromacs.org_gmx-users mailing list