Thank you for this. I tried csg_boltmann and got potentials and distributions for my bonds and angles. I tried to compare those distributions with those obtained by Gromacs using g_rdf -rdf mol_com and the equilibrium distance (maximum of distributions on x axis) lies in the same place but distribution obtained from gromacs is thousands times taller. Could you please explain this?
Can csg_boltmann do dihedrals as well? Or would you recomment g_stat for this? Steven W dniu piątek, 8 marca 2013 17:21:31 UTC użytkownik Christoph Junghans napisał: > > 2013/3/8 <[email protected] <javascript:>>: > > > > > > W dniu poniedziałek, 4 marca 2013 18:02:23 UTC użytkownik Christoph > Junghans > > napisał: > >> > >> Hi Steven, > >> > >> 2013/3/4 <[email protected]>: > >> > Dear Votca Users, > >> > > >> > I am new to the software. I tried some tutorials but many things are > >> > still > >> > not clear for me as my system is slightly more complex. > >> Welcome. > > > > > > Thank you for your prompt reply. > > > >> > >> > > >> > My atomistic simulation is one trimer of 3 amino acids in TIP3P water > >> > model. > >> > What I want to extract is just the bonded parameters by IBI between 3 > >> > beads > >> > ech represented by amino acid. > >> > > >> > So the key is the tabulated potential of two bonded interactions and > one > >> > angle. My questions: > >> > > >> > Does my reference simulation (atomistic) obtained by Gromacs should > have > >> > [ > >> > exlusions ] of non-bonded parameters so the potentials will be given > >> > only > >> > for bonded? Shall I include in my mdp file while performing > atomistic: > >> > energy_grps so I can use g_energy then to see the bonded potential? I > >> > know > >> > VOTCA does this but not sure based on what kind of trajectory (with > >> > excluded > >> > nonbonded?). > >> Basically there are two strategies to obtain coarse-grained potential > >> for bonded interactions. > >> 1.) Separation of bonded and non-bonded. (see Tschoep et al. Acta > >> Polymerica 49, 61 (1998).) > >> - run the molecule in vacuum (without solvent) and exclude all > >> interaction, which are later covered by non-bonded interactions. > >> - use simple Boltzmann inversion to obtain the bonded interaction (use > >> csg_stat/awk or csg_boltzmann for that) > > > > > > Shall include [ exclusions ] in my atomistic simulation on vacuum of all > > atoms from the desired 1st bead with atoms from the 2nd and 3rd bead + > > exlclusions of atoms of the 2nd bead with atms of the 3rd? Or its better > to > > group atoms in my index file in GMX - Bead1, Bead2, Bead3 and then > > in my mdp run: > > energygrp-excl = Bead1 Bead2 Bead1 Bead3 Bead2 Bead3 > Energy group exclusions are on top of the topology, Gromacs has > another loop over all pairs of energy group. > And because they are not part of the molecular topology, VOTCA don't > know about them. > > But these exclusion would only matter if you calculate distribution > for atomic (unmapped) coordinated using csg_stat > As long the bead were somehow excluded during atomistic simulation and > are excluded after the mapping (csg_dump --cg --excl ) > everything should be fine. > > > ?? How do you personally do it? > I usually put things in [exclusion], but that is more my personal > preference. Gromacs gets slower the more energy group you have, but > that is not so important for a single molecule in vacuum. > > > > > >> - obtain the other interaction by IBI of a cg simulation with the > >> above bonded interactions fixed (use csg_inverse) > >> 2.) Do IBI on all interaction (use csg_inverse) > >> - due to our lazyness in the implementation, you will also need > >> initial guesses for the bonded table here as well and hence you have > >> to the 2nd step of 1.). > >> > >> If 1.) or 2.) is better depends a bit on how stiff your bonded > >> interactions are and hence how good the separation works > > > > > > In this case I think it is good to do the step 1 and then the table of > > bonded can be an input to the second one so results will be more > realistic. > > Am I right? > In 2.) the tables for the bonded interaction are used as initial > guesses, an IBI update will still be applied on these, but it always > good to try 1.) first as it saves you a lot of time and work if it > works. > > > > > >> > >> > >> Concerning the exclusions, all VOTCA programs tolerate the exclusion > >> given in the tpr file if you don't map. > >> If you map first, VOTCA generates its own exclusion, which basically > >> mean every pair which is connected by a bonded interaction (specified > >> in the mapping file) is excluded. In both cases, exclusions can be > >> check using csg_dump --excl (together with --cg for mapped > >> trajectories). > >> > >> > > >> > Presuming I got my 3 bonded potentials (table.xvg). Shall i use in my > CG > >> > run: > >> > > >> > mdrun -v -tablep tablep1.xvg tablep2.xvg -tablea tablea.xvg -deffnm > md > >> > > >> > How Gromacs know which potential belong to which bond? Isnt it easier > to > >> > fit > >> > the curve into the parabolic shape and extract both equilibrium > >> > distance/angle and the force constant adding to ffbonded.itp ? > >> Gromacs has a naming scheme for its tables. > >> For bonded tables: > >> <TRUNC>_<A,B or D><NR>.xvg > >> and for non-bonded tables: > >> <TRUNC>_<E1><_E2>.xvg > >> > >> TRUNC is determined by the -table (nonbonded) and -tableb (bonded) > >> options. > >> E1,E2 are energy groups > >> A,B,D stands for angle, bond, dihedral > >> NR is the integer you specify in 4th column for of the bonds or angles > >> block in the top file. > >> For example > >> [ angles ] > >> ; ai aj ak funct NR k > >> 1 2 3 8 1 1.0 > >> (k ist another scale factor) > >> > >> I normally just use the default naming convention (mdrun without > >> -table options), which lead to table names like > >> table_A_B.xvg, table_b1.xvg and table_b1.xvg. > >> > >> Christoph > >> > >> > > >> > I am sorry for basic questions but it is not clear to me. > >> > > >> > Thanks, > >> > > >> > > >> > Steven > >> > > >> > > >> > > >> > -- > >> > You received this message because you are subscribed to the Google > >> > Groups > >> > "votca" group. > >> > To unsubscribe from this group and stop receiving emails from it, > send > >> > an > >> > email to [email protected]. > >> > To post to this group, send email to [email protected]. > >> > Visit this group at http://groups.google.com/group/votca?hl=en. > >> > For more options, visit https://groups.google.com/groups/opt_out. > >> > > >> > > >> > >> > >> > >> -- > >> Christoph Junghans > >> Web: http://www.compphys.de > > > > -- > > You received this message because you are subscribed to the Google > Groups > > "votca" group. > > To unsubscribe from this group and stop receiving emails from it, send > an > > email to [email protected] <javascript:>. > > To post to this group, send email to [email protected]<javascript:>. > > > Visit this group at http://groups.google.com/group/votca?hl=en. > > For more options, visit https://groups.google.com/groups/opt_out. > > > > > > > > -- > Christoph Junghans > Web: http://www.compphys.de > -- You received this message because you are subscribed to the Google Groups "votca" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/votca?hl=en. For more options, visit https://groups.google.com/groups/opt_out.
