Actually I would scale the potential, because for angles U=-kbt*log(dist/sin(angle)) and this normalization is done by csg_boltzmann.
For the normalizations, see eq. 5 of the VOTCA paper (<http://dx.doi.org/10.1021/ct900369w>) Cheers, Christoph 2011/6/27 chemistry <donets1...@gmail.com>: > In this case I have to multiply x value (rad) on (180/3.14) and > afterwords I can use this distribution function for getting the > potential and if it's not zero just simply shift it to zero? > > On Jun 27, 3:20 pm, Christoph Junghans <jungh...@votca.org> wrote: >> You can shift the minimum to zero, it will not change the simulation >> result (except for a energy offset). >> In VOTCA 1.2 the shifting is done automatically if you are using >> 'csg_call convert_potential gromacs' >> >> Notice that for angles csg_boltzmann calculates everything in rad, >> while gromacs (xvg files) wants degrees, so double check the scale of >> the x values again. >> >> Cheers, >> >> Christoph >> >> 2011/6/27 chemistry <donets1...@gmail.com>: >> >> > I calculated the potential with awk and with csg_boltzmann. It gives >> > the same shape of the curves but in the case when I used awk the >> > minimum of the potential has negative value. Should I shift the >> > minimum of the potential to zero? Does csg_boltzmann do it >> > automatically? >> >> > Thanks a lot in advance, >> > Sergio >> >> > On Jun 26, 4:42 pm, Christoph Junghans <jungh...@votca.org> wrote: >> >> Hi Sergio, >> >> >> the potentials have to look similar around the minimum. The region >> >> left and right are usually extrapolated and the exact form of the >> >> extrapolation does not matter much as the appearance of these values >> >> is exponentially suppressed (exp -E/k_bT). >> >> Make sure that the area {U_min..U_min+k_b T} is correct. >> >> >> Technically what I do to obtain table_b1.xvg is: >> >> -calculate the distribution with csg_stat / csg_boltzmann >> >> -calculate the potential using awk >> >> $ awk -v kbt=NUMBER '{print $1,($1>0)?-kbt*log($2/$1/$1):"nan"}' >> >> bond.dist.new > bond.pot.new >> >> -take a look at the potential and cut it down to the well-sampled region >> >> -convert it to gromacs format (this was reworked in VOTCA 1.2, before >> >> the extrapolation had to be done by hand) >> >> $ csg_call --options bond.xml --ia-type bonded convert_potential >> >> gromacs bond.pot.new2 table_b1.xvg >> >> where bond.xml looks like this: >> >> <cg> >> >> <inverse> >> >> <program>gromacs</program> >> >> <gromacs> >> >> <pot_max>1e8</pot_max> >> >> <table_end>3</table_end> >> >> <table_bins>0.002</table_bins> >> >> </gromacs> >> >> </inverse> >> >> </cg> >> >> >> table.xvg is used to calculate the interactions for all pairs, which >> >> do not appear in energygrp_table (in the mdp file). >> >> If all pairs are covered by energygrp_table combinations you can use >> >> any table, but I would use a table filled with zeros. >> >> table.xvg can be useful in cases when you want to mix IBI tables with >> >> a standard force field, then table.xvg should be the default 6-12 LJ >> >> table ($GMXDATA/gromacs/top/table6-12.xvg) >> >> >> Cheers, >> >> >> Christoph >> >> >> 2011/6/26 chemistry <donets1...@gmail.com>: >> >> >> > Thanks a lot for the previous advices! >> >> >> > I have one more question about table_*.xvg. Can you explain please how >> >> > exactly did you create all tables for the propane which you used for >> >> > the ibi-procedure (table.xvg, table_a1.xvg, table_b1.xvg). I did the >> >> > post-processing of the potentials (bond and angle) and I've got the >> >> > same shape for the potentials but much more higher value for the y- >> >> > axis. The distribution functions and potentials for bond and angle are >> >> > the same like in your presentation. Did I do something wrong or is it >> >> > something else what I should know? And which role plays table.xvg in >> >> > this case? >> >> >> > Thanks a lot in advance, >> >> > Sergio. >> >> >> > On Jun 23, 11:34 am, Christoph Junghans <jungh...@votca.org> wrote: >> >> >> Hi Sergio, >> >> >> >> 2011/6/22 chemistry <donets1...@gmail.com>:> Thanks a lot for advice, >> >> >> I checked all again and I found the mistake >> >> >> > in my mappin.xml file. I fixed it and now it works well. >> >> >> > Is it something important in the calculation of the rdf in the >> >> >> > iterations for such a systems? >> >> >> >> Usually one don't want to consider pairs of beads, which interact via >> >> >> a bonded potential, for the rdf. >> >> >> The reason is that the rdf is used to calculate the non-bonded >> >> >> interaction and so only the pairs which interact non-bonded should go >> >> >> in the rdf. >> >> >> >> However in some cases it is useful to calculate the rdf with different >> >> >> exclusions than the simulation was performed. (the >> >> >> cg.inverse.gromacs.rdf.topol option in the xml setting file, see the >> >> >> manual) >> >> >> >> Cheers, >> >> >> >> Christoph >> >> >> >> > Thanks a lot in advance. >> >> >> > Sergio >> >> >> >> > On Jun 21, 6:02 pm, Victor Ruehle <rue...@votca.org> wrote: >> >> >> >> Hmm, are you sure that all the beads which should be excluded are >> >> >> >> part of an angle definition? If yes, Votca should exclude them >> >> >> >> automatically.. >> >> >> >> >> I'm doing a very similar thing and it worked for me. Will check >> >> >> >> tomorrow whether there might be a new bug.However, third or fourth >> >> >> >> nearest neighbors should still be visible, there is no option so far >> >> >> >> to exclude all beads in the molecule. I could add that if necessary. >> >> >> >> >> Please verify if the peaks you observe really correspond to second >> >> >> >> nearest neighbors, and if yes whether you have this particular >> >> >> >> angle >> >> >> >> defined in the mapping. Also check csg_dump --excl >> >> >> >> >> 2011/6/21 chemistry <donets1...@gmail.com>: >> >> >> >> >> > I'm trying to calculate the target rdf for polymer system with the >> >> >> >> > "csg_stat" and also I'm interested to calculate the rdf in the >> >> >> >> > iterations: >> >> >> >> >> > csg_stat --top topol.tpr --trj traj.xtc --options settings.xml >> >> >> >> > --cg >> >> >> >> > mapping.xml >> >> >> >> >> > I have P3HT system. Every monomer divided on the three beads. >> >> >> >> > Thiophene ring represented the first bead and hexyl chain divided >> >> >> >> > on >> >> >> >> > the two beads - each bead include three carbon atoms. >> >> >> >> >> > in the settings.xml file I specified the types of the beads which >> >> >> >> > are >> >> >> >> > the same like in the mapping.xml: >> >> >> >> >> > ............. >> >> >> >> > <non-bonded> >> >> >> >> > <!-- name of the interaction --> >> >> >> >> > <name>P1-P1</name> >> >> >> >> > <!-- types involved in this interaction --> >> >> >> >> > <type1>A</type1> >> >> >> >> > <type2>A</type2> >> >> >> >> > <!-- dimension + grid spacing of tables for calculations --> >> >> >> >> > <min>0</min> >> >> >> >> > <max>1.36</max> >> >> >> >> > <step>0.01</step> >> >> >> >> > <inverse> >> >> >> >> > <!-- target distribution (rdf), just give gromas rdf.xvg --> >> >> >> >> > <target>P1-P1.dist.tgt</target> >> >> >> >> > <!-- update cycles --> >> >> >> >> > <do_potential>1 0 0</do_potential> >> >> >> >> > <!-- additional post processing of dU before added to >> >> >> >> > potential >> >> >> >> > --> >> >> >> >> > <post_update></post_update> >> >> >> >> > <!-- additional post processing of U after dU added to >> >> >> >> > potential >> >> >> >> > --> >> >> >> >> > <post_add></post_add> >> >> >> >> > <!-- name of the table for gromacs run --> >> >> >> >> > <gromacs> >> >> >> >> > <table>table_P1_P1.xvg</table> >> >> >> >> > </gromacs> >> >> >> >> > </inverse> >> >> >> >> > </non-bonded> >> >> >> >> > ............... >> >> >> >> >> > ---mapping.xml---: >> >> >> >> >> > <cg_molecule> >> >> >> >> > <name>P3HT</name> >> >> >> >> > <ident>Protein_X</ident> >> >> >> >> > <topology> >> >> >> >> > <cg_beads> >> >> >> >> > <cg_bead> >> >> >> >> > <name>S1</name> >> >> >> >> > <type>A</type> >> >> >> >> > <symmetry>1</symmetry> >> >> >> >> > <mapping>A</mapping> >> >> >> >> > <beads> 1:BBB:S01 1:BBB:C01 1:BBB:C02 1:BBB:H01 1:BBB:C03 >> >> >> >> > 1:BBB:C04 1:BBB:H02 </beads> >> >> >> >> > </cg_bead> >> >> >> >> > ................ >> >> >> >> >> > Also I specified all <bonds> and <angles> and <weights>. It works >> >> >> >> > well, but the second nearest neighbor does not excluded from the >> >> >> >> > rdf. >> >> >> >> > Seems that the exclusion works just for the first nearest >> >> >> >> > neighbor. Is >> >> >> >> > it possible to exclude this interaction from the rdf? >> >> >> >> >> > On Jun 17, 11:11 am, Victor Rühle <rue...@votca.org> wrote: >> >> >> >> >> Hey, >> >> >> >> >> >> Can you please describe you problem in more details, what >> >> >> >> >> exactly did >> >> >> >> >> you try? Are you referring to calculating the target rdf, or the >> >> >> >> >> rdf in >> >> >> >> >> the iterations? >> >> >> >> >> >> We introduced proper exclusion handling in version 1.1, so it >> >> >> >> >> should >> >> >> >> >> work. You can check whether exclusions are read correctly by: >> >> >> >> >> >> csg_dump --top topol.tpr --excl >> >> >> >> >> >> or if you use csg_stat to calculate target rdf: >> >> >> >> >> >> csg_dump --top topol.tpr --cg mappin.xml --excl >> >> >> >> >> >> If you use g_rdf, you have to specify the topology -s topol.tpr >> >> >> >> >> in >> >> >> >> >> addition to an index file. >> >> >> >> >> >> A brief side-note: >> >> >> >> >> Be aware, that for the reference rdf and that during the >> >> >> >> >> iterations, the >> >> >> >> >> exact identical normalization should be used (there is room for >> >> >> >> >> some >> >> >> >> >> discrepancy, however as far as i could see it's more problematic >> >> >> >> >> for IMC >> >> >> >> >> than for IBI). Therefore better use the same program for both >> >> >> >> >> (csg_stat >> >> >> >> >> to calculate reference). In version 1.2, which was released >> >> >> >> >> yesterday, >> >> >> >> >> we have adjusted the normalization to be identical to g_rdf in >> >> >> >> >> gromacs 4.5. >> >> >> >> >> >> Victor >> >> >> >> >> >> On 06/17/2011 10:32 AM, chemistry wrote: >> >> >> >> >> >> > I have one addition question about the ibi procedure. I have a >> >> >> >> >> > polymer >> >> >> >> >> > chain and I'm doing ibi for non-bonded interactions. How can I >> >> >> >> >> > exclude >> >> >> >> >> > the bond part from the RDF? Because in my case I have all the >> >> >> >> >> > same >> >> >> >> >> > beads in the backbone chain of my polymer and when I'm >> >> >> >> >> > calculating the >> >> >> >> >> > RDF it does not correspond to the purely non-bond >> >> >> >> >> > interactions. This >> >> >> >> >> > RDF except the non-bonded part includes the bonded part. >> >> >> >> >> >> > Thanks a lot in advance. >> >> >> >> >> > Sergio >> >> >> >> >> >> > On Jun 14, 2:59 pm, chemistry <donets1...@gmail.com> wrote: >> >> >> >> >> >> Thanks a lot for advice. Now ibi procedure works. >> >> >> >> >> >> >> On Jun 9, 8:31 pm, Victor Ruehle <rue...@votca.org> wrote: >> >> >> >> >> >> >>> Hey, >> >> >> >> >> >>> did you source your VOTCA installation? >> >> ... >> >> read more » > > -- > You received this message because you are subscribed to the Google Groups > "votca" group. > To post to this group, send email to votca@googlegroups.com. > To unsubscribe from this group, send email to > votca+unsubscr...@googlegroups.com. > For more options, visit this group at > http://groups.google.com/group/votca?hl=en. > > -- Dr. Christoph Junghans Votca Core Developer Web: http://www.votca.org -- You received this message because you are subscribed to the Google Groups "votca" group. To post to this group, send email to votca@googlegroups.com. To unsubscribe from this group, send email to votca+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/votca?hl=en.
