<div>Dear Gmx Users, Dear Justin,</div>
<div> </div>
<div>I pulled my ligand away from my protein. Ligand was attached to lower part of my protein, I pulled in Z coordinate it using:</div>
<div><span lang="EN">
<p>; Run parameters</p>
<p>integrator = md ; leap-frog integrator</p>
<p>nsteps = 5000000 ; 2 * 5000000 = 10 ns</p>
<p>dt = 0.002 ; 2 fs</p>
<p>tinit = 0</p>
<p>nstcomm = 10</p>
<p>; Output control</p>
<p>nstxout = 50000 ; save coordinates every 100 ps</p>
<p>nstvout = 50000 ; save velocities every </p>
<p>nstfout = 5000</p>
<p>nstxtcout = 5000 ; every 10 ps</p>
<p>nstenergy = 5000</p>
<p>; Bond parameters</p>
<p>continuation = yes ; first dynamics run</p>
<p>constraint_algorithm = lincs ; holonomic constraints </p>
<p>constraints = all-bonds ; all bonds (even heavy atom-H bonds) constrained</p>
<p>; Neighborsearching</p>
<p>ns_type = grid ; search neighboring grid cells</p>
<p>nstlist = 5 ; 10 fs</p>
<p>rlist = 0.9 ; short-range neighborlist cutoff (in nm)</p>
<p>rcoulomb = 0.9 ; short-range electrostatic cutoff (in nm)</p>
<p>rvdw = 0.9 ; short-range van der Waals cutoff (in nm)</p>
<p>ewald_rtol = 1e-5 ; relative strenght of the Ewald-shifted potential rcoulomb</p>
<p>; Electrostatics</p>
<p>coulombtype = PME ; Particle Mesh Ewald for long-range electrostatics</p>
<p>pme_order = 4 ; cubic interpolation</p>
<p>fourierspacing = 0.12 ; grid spacing for FFT</p>
<p>fourier_nx = 0</p>
<p>fourier_ny = 0</p>
<p>fourier_nz = 0</p>
<p>optimize_fft = yes</p>
<p>; Temperature coupling is on</p>
<p>tcoupl = V-rescale ; modified Berendsen thermostat</p>
<p>tc_grps = Protein_LIG Water_and_ions ; two coupling groups - more accurate</p>
<p>tau_t = 0.1 0.1 ; time constant, in ps</p>
<p>ref_t = 298 298 ; reference temperature, one for each group, in K</p>
<p>; Pressure coupling is on</p>
<p>pcoupl = Parrinello-Rahman ; pressure coupling is on for NPT</p>
<p>pcoupltype = isotropic ; uniform scaling of box vectors</p>
<p>tau_p = 1.0 ; time constant, in ps</p>
<p>ref_p = 1.0 ; reference pressure, in bar</p>
<p>compressibility = 4.5e-5 ; isothermal compressibility of water, bar^-1</p>
<p>; Periodic boundary conditions</p>
<p>pbc = xyz ; 3-D PBC</p>
<p>; Dispersion correction</p>
<p>DispCorr = EnerPres ; account for cut-off vdW scheme</p>
<p>; Velocity generation</p>
<p>gen_vel = no ; assign velocities from Maxwell distribution</p>
<p>; These options remove COM motion of the system</p>
<p>; Pull code</p>
<p>pull = umbrella</p>
<p>pull_geometry = distance</p>
<p>pull_dim = N N Y</p>
<p>pull_start = yes </p>
<p>pull_ngroups = 1</p>
<p>pull_group0 = Protein </p>
<p>pull_group1 = LIG182 </p>
<p>pull_init1 = 0</p>
<p>pull_rate1 = 0.0</p>
<p>pull_k1 = 200 ; kJ mol^-1 nm^-2</p>
<p>pull_nstxout = 1000 ; every 2 ps</p>
<p>pull_nstfout = 1000 ; every 2 ps</p>
<p>Following Justin's tutorial I used perl script to extract coordinate for each window. </p></span></div>
<div><span lang="EN">
<p>0 2.4595039</p>
<p>1 2.4745028</p>
<p>...</p>
<p>500 8.74</p>
<p>My ligand at the begining was at such distance as it was in the lower part of the protein. Then I used 0.1 nm spacing at the begining (till 4 nm) and 0.2 nm later on.</p>
<p>And following equilibration in each window I run umbrella sampling for 10ns in app 49 windows:</p><span lang="EN">
<p>Run parameters</p>
<p>integrator = md ; leap-frog integrator</p>
<p>nsteps = 5000000 ; 2 * 5000000 = 10 ns</p>
<p>dt = 0.002 ; 2 fs</p>
<p>tinit = 0</p>
<p>nstcomm = 10</p>
<p>; Output control</p>
<p>nstxout = 50000 ; save coordinates every 100 ps</p>
<p>nstvout = 50000 ; save velocities every </p>
<p>nstfout = 5000</p>
<p>nstxtcout = 5000 ; every 10 ps</p>
<p>nstenergy = 5000</p>
<p>; Bond parameters</p>
<p>continuation = yes ; first dynamics run</p>
<p>constraint_algorithm = lincs ; holonomic constraints </p>
<p>constraints = all-bonds ; all bonds (even heavy atom-H bonds) constrained</p>
<p>; Neighborsearching</p>
<p>ns_type = grid ; search neighboring grid cells</p>
<p>nstlist = 5 ; 10 fs</p>
<p>rlist = 0.9 ; short-range neighborlist cutoff (in nm)</p>
<p>rcoulomb = 0.9 ; short-range electrostatic cutoff (in nm)</p>
<p>rvdw = 0.9 ; short-range van der Waals cutoff (in nm)</p>
<p>ewald_rtol = 1e-5 ; relative strenght of the Ewald-shifted potential rcoulomb</p>
<p>; Electrostatics</p>
<p>coulombtype = PME ; Particle Mesh Ewald for long-range electrostatics</p>
<p>pme_order = 4 ; cubic interpolation</p>
<p>fourierspacing = 0.12 ; grid spacing for FFT</p>
<p>fourier_nx = 0</p>
<p>fourier_ny = 0</p>
<p>fourier_nz = 0</p>
<p>optimize_fft = yes</p>
<p>; Temperature coupling is on</p>
<p>tcoupl = V-rescale ; modified Berendsen thermostat</p>
<p>tc_grps = Protein_LIG Water_and_ions ; two coupling groups - more accurate</p>
<p>tau_t = 0.1 0.1 ; time constant, in ps</p>
<p>ref_t = 298 298 ; reference temperature, one for each group, in K</p>
<p>; Pressure coupling is on</p>
<p>pcoupl = Parrinello-Rahman ; pressure coupling is on for NPT</p>
<p>pcoupltype = isotropic ; uniform scaling of box vectors</p>
<p>tau_p = 1.0 ; time constant, in ps</p>
<p>ref_p = 1.0 ; reference pressure, in bar</p>
<p>compressibility = 4.5e-5 ; isothermal compressibility of water, bar^-1</p>
<p>; Periodic boundary conditions</p>
<p>pbc = xyz ; 3-D PBC</p>
<p>; Dispersion correction</p>
<p>DispCorr = EnerPres ; account for cut-off vdW scheme</p>
<p>; Velocity generation</p>
<p>gen_vel = no ; assign velocities from Maxwell distribution</p>
<p>; These options remove COM motion of the system</p>
<p>; Pull code</p>
<p>pull = umbrella</p>
<p>pull_geometry = distance</p>
<p>pull_dim = N N Y</p>
<p>pull_start = yes </p>
<p>pull_ngroups = 1</p>
<p>pull_group0 = Protein </p>
<p>pull_group1 = LIG182 </p>
<p>pull_init1 = 0</p>
<p>pull_rate1 = 0.0</p>
<p>pull_k1 = 200 ; kJ mol^-1 nm^-2</p>
<p>pull_nstxout = 1000 ; every 2 ps</p>
<p>pull_nstfout = 1000 ; every 2 ps</p></span>
<p> </p>
<p>My PMF profile:</p>
<p><a href="http://speedy.sh/zerqZ/profile.JPG">http://speedy.sh/zerqZ/profile.JPG</a></p>
<p>My histogram: <a href="http://speedy.sh/PyhnN/Histo.JPG">http://speedy.sh/PyhnN/Histo.JPG</a></p>
<p>Why g_wham takes into account distances below 2.45 nm as the 1st structure was at 2.45. If I get rid of the distances below 2.45 (those weird values PMF values) I obtain beautiful profile:</p>
<p><a href="http://speedy.sh/TUXGC/profile1.JPG">http://speedy.sh/TUXGC/profile1.JPG</a></p>
<p>Please, explain!</p>
<p>Steven</p>
<p> </p></span></div>