On Tue, Dec 18, 2018 at 8:44 PM Lee <liuchongm...@gmail.com> wrote: > > I have fitted the distribution by Gaussian curve (the red line) and > use it as the target distribution. But the converted potential presents not a > harmonic form (as shown in ACF.pot.cur.png). I want to know what causes the > potential curve drops sharply at the boundary PI. > And I don't konw what's the meaning of "use an angle type 1 (or 6) in > the gromacs topology for that angle interaction". The MD simulation is > performed by lammps. In my topology, this angle type is already defined.
Sorry, what I meant to say was that you could just use an quadratic angle with the parameters you got from the fit and don't even iterate that interaction at all. Christoph > > 在 2018年12月19日星期三 UTC+8上午12:01:14，Christoph Junghans写道： >> >> On Tue, Dec 18, 2018 at 6:14 AM Lee <liucho...@gmail.com> wrote: >> > >> > Hi Christoph, >> > Now I want to fit the target distribution with single peak gaussian >> > distribution. So the angle potential can be expressed with harmonic >> > potential (quadratic form). But if the peak of angle distribution close to >> > PI, the target distribution must be truncated at the boundary (PI) with a >> > big probability density value (as shown in the figure attched). The angle >> > potential converted from this target potential doesn't show a harmonic >> > shape. The right part of potential close to PI drops sharply to zero. >> > Theoretically, it should present as standard quadratic form just like the >> > left half part. So I doubt whether the truncation is the reason because >> > the left part without truncated by boundary is perfect. >> > How can I get a standard harmonic potential in this situation? Or >> > maybe it's impossible. >> >> Just fit the distribution by hand and use an angle type 1 (or 6) in >> the gromacs topology for that angle interaction. >> >> Christoph >> >> > Thanks very much >> > >> > 在 2018年12月7日星期五 UTC+8下午8:56:34，Christoph Junghans写道： >> >> >> >> On Fri, Dec 7, 2018 at 4:54 AM Lee <liucho...@gmail.com> wrote: >> >> > >> >> > After I added the pressure correction, the density of the cg system is >> >> > very close to the full-atomic system. Thanks for your advise. But when >> >> > I use these final potential table to do the tensile simulation for cg >> >> > system, the system is unstable. And the stress-strain curve is extrmely >> >> > terrible. I saw someone abtained a good stress-strain curve at >> >> > https://groups.google.com/forum/#!msg/votca/Jxsdz6v9QKw/OUVfYan8CwAJ . >> >> > So I want to know how could I get a good stress-strain curve with these >> >> > potential table? Or there still need some post process for these table? >> >> >> >> That is magic of coarse-graining, if you parameterize a model on RDF >> >> and pressure, it isn't obvious that the model will reproduce the >> >> stress-strain curve. Maybe Frank knows some details and can comment >> >> here. >> >> >> >> Christoph >> >> > >> >> > 在 2018年12月3日星期一 UTC+8下午10:58:40，Christoph Junghans写道： >> >> >> >> >> >> On Mon, Dec 3, 2018 at 3:51 AM Lee <liucho...@gmail.com> wrote: >> >> >> > >> >> >> > Yes, you are right. I missed a important item, pressure >> >> >> > correction. But when I add pressure into the <post_update>, the >> >> >> > error occurred "Pressure correction for lammps is not implemented >> >> >> > yet". Does the pressure correction need to be implemented >> >> >> > additionally? And how to implement it? >> >> >> >> >> >> Well it just means we haven't added support for pressure correction >> >> >> when using lammps in that version. >> >> >> However, it has been added here https://github.com/votca/csg/pull/201, >> >> >> so either patch that in or use the development version of votca >> >> >> instead (see >> >> >> https://github.com/votca/votca/blob/master/share/doc/INSTALL.md#general-source-installation-instructions.) >> >> >> >> >> >> Christoph >> >> >> >> >> >> > Thanks >> >> >> > >> >> >> > 在 2018年12月3日星期一 UTC+8上午10:48:57，Christoph Junghans写道： >> >> >> >> >> >> >> >> On Sun, Dec 2, 2018 at 6:40 PM Lee <liucho...@gmail.com> wrote: >> >> >> >> > >> >> >> >> > By checking the convergence value of interaction, I found that it >> >> >> >> > was stabilized above 1. When I use these potential table to do >> >> >> >> > the cg simulation, the density of cg system is half of that the >> >> >> >> > corresponding atomic system. So I think the nonbond interaction >> >> >> >> > is still immature. How to reach a lower convergence number? The >> >> >> >> > rdf and the convergence number of C-C are attched. >> >> >> >> >> >> >> >> Run longer iterations, finer binsize, in general everything which >> >> >> >> increases the statistics helps. >> >> >> >> >> >> >> >> On the other hand, your RDF conversion looks pretty good already, >> >> >> >> don't let this one number guide you, you need to test the >> >> >> >> coarse-grained model, too. >> >> >> >> >> >> >> >> Christoph >> >> >> >> > >> >> >> >> > 在 2018年12月1日星期六 UTC+8下午9:44:34，Christoph Junghans写道： >> >> >> >> >> >> >> >> >> >> On Fri, Nov 30, 2018 at 8:00 PM Lee <liucho...@gmail.com> wrote: >> >> >> >> >> > >> >> >> >> >> > OK, thanks! I have increased the bin size and the >> >> >> >> >> > distribution seems to be smoothed. >> >> >> >> >> > In some paper, the IBI process is started from bond and >> >> >> >> >> > then angle, the nonbonded potential is finished at last. When >> >> >> >> >> > doing ibi for bond and angle, the pair interaction is turned >> >> >> >> >> > off. The chains are pushed apart to remove close contacts by >> >> >> >> >> > slowly turning on a soft, purely repulsive potential as you >> >> >> >> >> > said before. But I couldn't find the command in lammps. >> >> >> >> >> That might be a question for the lammps mailing list! >> >> >> >> >> >> >> >> >> >> > Another question, how could I output the value of >> >> >> >> >> > convergence for every iterative step? >> >> >> >> >> The convergence numbers are stores in a files for each >> >> >> >> >> interaction >> >> >> >> >> called <NAME>.conv and the sum is printed in the log files, so >> >> >> >> >> you can >> >> >> >> >> do: >> >> >> >> >> $ grep "Convergence sum" inverse.log >> >> >> >> >> >> >> >> >> >> This of course assumes you enabled the convergence check feature >> >> >> >> >> (with >> >> >> >> >> setting cg.inverse.convergence_check.type to "default") >> >> >> >> >> >> >> >> >> >> Christoph >> >> >> >> >> > >> >> >> >> >> > 在 2018年11月30日星期五 UTC+8下午11:24:33，Christoph Junghans写道： >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> On Fri, Nov 30, 2018 at 01:05 Lee <liucho...@gmail.com> wrote: >> >> >> >> >> >>> >> >> >> >> >> >>> In votca, the smoothing process works for potential. Is >> >> >> >> >> >>> there any smoothing process for new distribution? And what's >> >> >> >> >> >>> the command? >> >> >> >> >> >> >> >> >> >> >> >> No, but it would be hard to add! >> >> >> >> >> >> >> >> >> >> >> >> Also increasing the bin size naturally does smoothing! >> >> >> >> >> >> >> >> >> >> >> >> Christoph >> >> >> >> >> >>> >> >> >> >> >> >>> >> >> >> >> >> >>> >> >> >> >> >> >>> 在 2018年11月21日星期三 UTC+8下午9:03:19，Christoph Junghans写道： >> >> >> >> >> >>>> >> >> >> >> >> >>>> On Tue, Nov 20, 2018 at 6:44 PM Lee <liucho...@gmail.com> >> >> >> >> >> >>>> wrote: >> >> >> >> >> >>>> > >> >> >> >> >> >>>> > The expression of boltzmann inversion is V=-kBTln(P). And >> >> >> >> >> >>>> > the distribution P is related to the scale of the >> >> >> >> >> >>>> > x-coordinate. The plot of bond distribution with >> >> >> >> >> >>>> > different units are attached. In votca, all the processes >> >> >> >> >> >>>> > including calculating distributions and boltzmann >> >> >> >> >> >>>> > inversion are finished with units nm. So can I say that >> >> >> >> >> >>>> > the obtained potential table used in lammps is converted >> >> >> >> >> >>>> > by V=-kBTln(P) with the length units nm? So in lammps, >> >> >> >> >> >>>> > the dynamic simulation is finished with length units >> >> >> >> >> >>>> > angstroms but with the potential corresponding to units >> >> >> >> >> >>>> > nm. >> >> >> >> >> >>>> >> >> >> >> >> >>>> Well, V(x)= -kBTln(P(x)) and x can be a distance or an >> >> >> >> >> >>>> angle etc. V >> >> >> >> >> >>>> and P are always parameterized by the same variable in its >> >> >> >> >> >>>> units. Also >> >> >> >> >> >>>> P(x) is a distribution and hence is unit-less, plus the >> >> >> >> >> >>>> normalization >> >> >> >> >> >>>> of the distribution doesn't matter much as ln(a*P)=ln a + >> >> >> >> >> >>>> ln P and >> >> >> >> >> >>>> hence only leads to an irrelevant offset in the potential >> >> >> >> >> >>>> and doesn't >> >> >> >> >> >>>> matter to the forces. >> >> >> >> >> >>>> >> >> >> >> >> >>>> Christoph >> >> >> >> >> >>>> >> >> >> >> >> >>>> > >> >> >> >> >> >>>> > 在 2018年11月20日星期二 UTC+8下午11:43:59，Christoph Junghans写道： >> >> >> >> >> >>>> >> >> >> >> >> >> >>>> >> On Tue, Nov 20, 2018 at 2:04 AM Lee >> >> >> >> >> >>>> >> <liucho...@gmail.com> wrote: >> >> >> >> >> >>>> >> > >> >> >> >> >> >>>> >> > I have another question. If my units command is >> >> >> >> >> >>>> >> > real in lammps which the units of distance and energy >> >> >> >> >> >>>> >> > are angstroms and Kcal/mole respectively, what shoud I >> >> >> >> >> >>>> >> > change in the setting file in order to get the >> >> >> >> >> >>>> >> > potential table with right units? Now, the bond and >> >> >> >> >> >>>> >> > angle potential of the CG system are much lower than >> >> >> >> >> >>>> >> > the corresponding atomic system. >> >> >> >> >> >>>> >> >> >> >> >> >> >>>> >> Usually VOTCA calculated things in Gromacs units (so nm, >> >> >> >> >> >>>> >> kJ/mol etc.), >> >> >> >> >> >>>> >> but if you only calculate RDFs you don't need to change >> >> >> >> >> >>>> >> anything as >> >> >> >> >> >>>> >> input and output are both in nm and hence there is no >> >> >> >> >> >>>> >> scaling >> >> >> >> >> >>>> >> involved. So this will work for angstroms as well. >> >> >> >> >> >>>> >> Additionally, you >> >> >> >> >> >>>> >> need to make sure that kBT is in the right unit in the >> >> >> >> >> >>>> >> settings file. >> >> >> >> >> >>>> >> >> >> >> >> >> >>>> >> Christoph >> >> >> >> >> >>>> >> > Thanks a lot. >> >> >> >> >> >>>> >> > >> >> >> >> >> >>>> >> > 在 2018年11月20日星期二 UTC+8上午5:52:29，Christoph Junghans写道： >> >> >> >> >> >>>> >> >> >> >> >> >> >> >>>> >> >> On Mon, Nov 19, 2018 at 8:51 AM Christoph Junghans >> >> >> >> >> >>>> >> >> <jung...@votca.org> wrote: >> >> >> >> >> >>>> >> >> > >> >> >> >> >> >>>> >> >> > On Sun, Nov 18, 2018 at 8:31 PM Lee >> >> >> >> >> >>>> >> >> > <liucho...@gmail.com> wrote: >> >> >> >> >> >>>> >> >> > > >> >> >> >> >> >>>> >> >> > > For the dynamic simulations finished with lammps, >> >> >> >> >> >>>> >> >> > > should I set the exclusions? If excluding >> >> >> >> >> >>>> >> >> > > non-bonded interactions between atoms that are no >> >> >> >> >> >>>> >> >> > > further than 3 bonds away, does it need to >> >> >> >> >> >>>> >> >> > > excluding these atoms when statistical target rdf >> >> >> >> >> >>>> >> >> > > distributon? >> >> >> >> >> >>>> >> >> > > >> >> >> >> >> >>>> >> >> > Yes, for LAMMPS you will most likely need to create >> >> >> >> >> >>>> >> >> > rdf exclusion manually. >> >> >> >> >> >>>> >> >> > >> >> >> >> >> >>>> >> >> > You can check, what exclusion VOTCA see by running: >> >> >> >> >> >>>> >> >> > $ csg_dump --top <your_topology_file> --excl >> >> >> >> >> >>>> >> >> > >> >> >> >> >> >>>> >> >> > For advanced topology handling see, section 3.3 of >> >> >> >> >> >>>> >> >> > the manual, >> >> >> >> >> >>>> >> >> > http://doc.votca.org/manual.pdf >> >> >> >> >> >>>> >> >> Also Joshua might be able to say something about the >> >> >> >> >> >>>> >> >> new LAMMPS >> >> >> >> >> >>>> >> >> topology reader in csg and how exclusions are >> >> >> >> >> >>>> >> >> incorporated there. >> >> >> >> >> >>>> >> >> >> >> >> >> >> >>>> >> >> Christoph >> >> >> >> >> >>>> >> >> > >> >> >> >> >> >>>> >> >> > Christoph >> >> >> >> >> >>>> >> >> > > -- >> >> >> >> >> >>>> >> >> > > 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 >> >> >> >> >> >>>> >> >> > > votca+un...@googlegroups.com. >> >> >> >> >> >>>> >> >> > > To post to this group, send email to >> >> >> >> >> >>>> >> >> > > vo...@googlegroups.com. >> >> >> >> >> >>>> >> >> > > Visit this group at >> >> >> >> >> >>>> >> >> > > https://groups.google.com/group/votca. >> >> >> >> >> >>>> >> >> > > For more options, visit >> >> >> >> >> >>>> >> >> > > https://groups.google.com/d/optout. >> >> >> >> >> >>>> >> >> > >> >> >> >> >> >>>> >> >> > >> >> >> >> >> >>>> >> >> > >> >> >> >> >> >>>> >> >> > -- >> >> >> >> >> >>>> >> >> > Christoph Junghans >> >> >> >> >> >>>> >> >> > Web: http://www.compphys.de >> >> >> >> >> >>>> >> >> >> >> >> >> >> >>>> >> >> >> >> >> >> >> >>>> >> >> >> >> >> >> >> >>>> >> >> -- >> >> >> >> >> >>>> >> >> 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 >> >> >> >> >> >>>> >> > votca+un...@googlegroups.com. >> >> >> >> >> >>>> >> > To post to this group, send email to >> >> >> >> >> >>>> >> > vo...@googlegroups.com. >> >> >> >> >> >>>> >> > Visit this group at >> >> >> >> >> >>>> >> > https://groups.google.com/group/votca. >> >> >> >> >> >>>> >> > For more options, visit >> >> >> >> >> >>>> >> > https://groups.google.com/d/optout. >> >> >> >> >> >>>> >> >> >> >> >> >> >>>> >> >> >> >> >> >> >>>> >> >> >> >> >> >> >>>> >> -- >> >> >> >> >> >>>> >> 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 votca+un...@googlegroups.com. >> >> >> >> >> >>>> > To post to this group, send email to >> >> >> >> >> >>>> > vo...@googlegroups.com. >> >> >> >> >> >>>> > Visit this group at https://groups.google.com/group/votca. >> >> >> >> >> >>>> > For more options, visit >> >> >> >> >> >>>> > https://groups.google.com/d/optout. >> >> >> >> >> >>>> >> >> >> >> >> >>>> >> >> >> >> >> >>>> >> >> >> >> >> >>>> -- >> >> >> >> >> >>>> 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 votca+un...@googlegroups.com. >> >> >> >> >> >>> To post to this group, send email to vo...@googlegroups.com. >> >> >> >> >> >>> Visit this group at https://groups.google.com/group/votca. >> >> >> >> >> >>> For more options, visit https://groups.google.com/d/optout. >> >> >> >> >> >> >> >> >> >> >> >> -- >> >> >> >> >> >> 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 votca+un...@googlegroups.com. >> >> >> >> >> > To post to this group, send email to vo...@googlegroups.com. >> >> >> >> >> > Visit this group at https://groups.google.com/group/votca. >> >> >> >> >> > For more options, visit https://groups.google.com/d/optout. >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> -- >> >> >> >> >> 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 votca+un...@googlegroups.com. >> >> >> >> > To post to this group, send email to vo...@googlegroups.com. >> >> >> >> > Visit this group at https://groups.google.com/group/votca. >> >> >> >> > For more options, visit https://groups.google.com/d/optout. >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> >> -- >> >> >> >> 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 votca+un...@googlegroups.com. >> >> >> > To post to this group, send email to vo...@googlegroups.com. >> >> >> > Visit this group at https://groups.google.com/group/votca. >> >> >> > For more options, visit https://groups.google.com/d/optout. >> >> >> >> >> >> >> >> >> >> >> >> -- >> >> >> 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 votca+un...@googlegroups.com. >> >> > To post to this group, send email to vo...@googlegroups.com. >> >> > Visit this group at https://groups.google.com/group/votca. >> >> > For more options, visit https://groups.google.com/d/optout. >> >> >> >> >> >> >> >> -- >> >> 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 votca+un...@googlegroups.com. >> > To post to this group, send email to vo...@googlegroups.com. >> > Visit this group at https://groups.google.com/group/votca. >> > For more options, visit https://groups.google.com/d/optout. >> >> >> >> -- >> 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 votca+unsubscr...@googlegroups.com. > To post to this group, send email to votca@googlegroups.com. > Visit this group at https://groups.google.com/group/votca. > For more options, visit https://groups.google.com/d/optout. -- 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 votca+unsubscr...@googlegroups.com. To post to this group, send email to votca@googlegroups.com. Visit this group at https://groups.google.com/group/votca. For more options, visit https://groups.google.com/d/optout.
