W dniu środa, 14 sierpnia 2013 15:19:12 UTC+1 użytkownik Christoph Junghans napisał: > > 2013/8/14 <[email protected] <javascript:>>: > > > > > > W dniu wtorek, 13 sierpnia 2013 17:37:21 UTC+1 użytkownik Christoph > Junghans > > napisał: > >> > >> 2013/8/13 <[email protected]>: > >> > > >> > > >> > W dniu wtorek, 13 sierpnia 2013 16:08:49 UTC+1 użytkownik Christoph > >> > Junghans > >> > napisał: > >> >> > >> >> 2013/8/13 <[email protected]>: > >> >> > > >> >> > > >> >> > W dniu wtorek, 13 sierpnia 2013 14:54:16 UTC+1 użytkownik > Christoph > >> >> > Junghans > >> >> > napisał: > >> >> >> > >> >> >> 2013/8/13 <[email protected]>: > >> >> >> > Or is there any command which will tell csg_stat to get rid of > PBC > >> >> >> > so > >> >> >> > that > >> >> >> > cell dimension will not have influcence on RDFs? I mean target > >> >> >> > distributions > >> >> >> > as well as during the iterations in IBI? > >> >> >> VOTCA reads the box from the tpr file. One could use a gro file > as a > >> >> >> topology (--top conf.gro), but the gro file contains less > >> >> >> information > >> >> >> (e.g. charges, mass, types will be wrong.). Gro files contain a > box, > >> >> >> too, but for some reason, which I don't remember right now, VOTCA > >> >> >> won't use it. > >> >> >> > >> >> >> You can use a different topology file to calculate the rdfs > >> >> >> (cg.inverse.gromacs.topol). We originally added this feature, so > >> >> >> that > >> >> >> one can calculated rdfs with a different number of exclusions, > but > >> >> >> one > >> >> >> can also use that to use a different box. In your case just make > the > >> >> >> box huge. > >> >> >> > >> >> >> Cheers, > >> >> >> > >> >> >> Christoph > > > > > > My atomistic simulation was run at NPT ensemble with a cubic box > dimension > > of 95A with fully explicit solvent. Now I do not have water as I want to > > include it in the non bonded parameters.Previously I had a box of 150A > in > > IBI (NVT - no water) and distributions looked like I attached in my 1st > post > > - Same shape but much higher (the box is bigger the bigger the > > distributions). Now I set up the box to 95A in and running IBI in NVT > and > > they have similar shape and values- they are converging. Is that a good > > approach? Or shall I use nopbc option in mdp as you suggested and try to > > converge it to atomisitic distributions without pbc as well? > > The box size doesn't appear anywhere in rdf calculation except for > treading pbc, so huge box and no pbc should do exactly the same thing. > > At this point I don't really understand where the different > normalization come from. Have you re-calculated the rdf in both case > (target and cg simulation) with csg_stat, g_rdf or vmd to see if it > is consistent? >
Thank you. csg_stat and g_rdf are consistent. VMD is hard to assess as I do not know how to apply exclusions. Anyhow I changed the size of the box with pbc to my reference atomistic simulation and they are converging really well. Once converged i will compare structural properties e.g. radius of gyration if that is the same. It is already really close to atomistic radius. Steven > > Concerning your approach, I am not the expert for implicit solvent > models, so I cannot judge, somebody else on the list might know more, > but using distribution from an NPT simulation as targets in an NVT > simulation might lead to systematic errors. > > > > > > Steven > > > > > >> > >> >> > > >> >> > > >> >> > Thank you. But when I set up the conf.gro to the box average from > >> >> > atomistic > >> >> > simulations the tpr file by grompp will be set up correctly. I > tried > >> >> > this > >> >> > and all profiles are converging. However, the more accurate will > be > >> >> > RDF > >> >> > which does not account the box size so without pbc. Is there any > >> >> > command > >> >> > in > >> >> > csg_stat which can calculate this like g_rdf -nopbc option? I > >> >> > realised > >> >> > in > >> >> > VMD that when changing the cell dimension RDFs changes....which > makes > >> >> > sense. > >> >> > When setup without box size (dont use pbc) I got normalized > >> >> > distributions > >> >> > which will be more accurate in my case. > >> >> No, csg_stat doesn't have such an option! (see "csg_stat --help" for > >> >> details). > >> >> However, feel free to add one ;-) > >> >> > >> >> Like I said before, the easiest in your case will be to create a > >> >> special tpr with no box beforehand and give it to csg_stat via > >> >> the cg.inverse.gromacs.topol option, so that the special tpr is only > >> >> used for the rdf calculation, but not for the simulation. > >> >> > >> >> Christoph > >> > > >> > > >> > Thank you. I guess grompp will complain about no box dimension but it > >> > can be > >> > done I guess with maxwarining option. > >> Or you could make the box huge like I suggested earlier. Also gromacs > >> has an mdp option pbc=no. > >> > >> Anyhow, this is more a gromacs question, please post it on their > mailing > >> list. > >> > >> > other would be to use g_rdf script you provided me, yep? > >> Maybe, but we removed that rdf script for several reasons, so there is > >> no official support for it. > >> If it doesn't work don't blame us ;-) > >> > >> Christoph > >> > > >> > Steven > >> >> > >> >> > > >> >> > Steven > >> >> > > >> >> >> > >> >> >> > > >> >> >> > Steven > >> >> >> > > >> >> >> > W dniu wtorek, 13 sierpnia 2013 11:34:51 UTC+1 użytkownik > >> >> >> > Valentina > >> >> >> > Erastova > >> >> >> > napisał: > >> >> >> >> > >> >> >> >> > >> >> >> >> On 13 Aug 2013, at 11:27, [email protected] wrote: > >> >> >> >> > >> >> >> >> > >> >> >> >> > >> >> >> >> W dniu wtorek, 13 sierpnia 2013 11:24:59 UTC+1 użytkownik > >> >> >> >> Valentina > >> >> >> >> Erastova napisał: > >> >> >> >>> > >> >> >> >>> > >> >> >> >>> On 13 Aug 2013, at 11:21, [email protected] wrote: > >> >> >> >>> > >> >> >> >>> I used different software ACEMD driving my protein from 300K > to > >> >> >> >>> 500K > >> >> >> >>> thousands of times to explore or possible conformations...so > the > >> >> >> >>> box > >> >> >> >>> dimension was changing from 9.4 to 11.4 nm. At 300K it had > 9.4 > >> >> >> >>> nm. > >> >> >> >>> I > >> >> >> >>> used > >> >> >> >>> all the states corresponding to 300K only. No I will change > the > >> >> >> >>> box > >> >> >> >>> in > >> >> >> >>> my CG > >> >> >> >>> iterative method to 9.4 nm as I had 12 nm no clue why... > Should > >> >> >> >>> work > >> >> >> >>> fine > >> >> >> >>> now. Thank you! > >> >> >> >>> > >> >> >> >>> > >> >> >> >>> May be some inconsistence with trajectory reading. > >> >> >> >>> Never used AceMD. > >> >> >> >>> V > >> >> >> >>> > >> >> >> >> > >> >> >> >> You better try that :) 60 ns a day with 100 K atoms using temp > >> >> >> >> annealing > >> >> >> >> ar one GPU :) > >> >> >> >> > >> >> >> >> > >> >> >> >> Just implemented material science FFs onto gromacs, not > >> >> >> >> reimplementing > >> >> >> >> again (pain). But sounds good;) > >> >> >> >> > >> >> >> >> > >> >> >> >>> > >> >> >> >>> W dniu wtorek, 13 sierpnia 2013 11:08:21 UTC+1 użytkownik > >> >> >> >>> Valentina > >> >> >> >>> Erastova napisał: > >> >> >> >>>> > >> >> >> >>>> So, you started with cubic box of 95A * 3 (gromacs units > are > >> >> >> >>>> nm)? > >> >> >> >>>> Then > >> >> >> >>>> you T annealed, got output confout.gro (with box size at the > >> >> >> >>>> bottom > >> >> >> >>>> of the > >> >> >> >>>> file), change of the box size can be taken from trajectory. > >> >> >> >>>> > >> >> >> >>>> I am not too sure about your method, tbh, as annealing is a > >> >> >> >>>> funny > >> >> >> >>>> thing > >> >> >> >>>> - it can cause system to lag behind. Personally, I wouldn't > use > >> >> >> >>>> a > >> >> >> >>>> system > >> >> >> >>>> that potentially can not have reached equilibrium to get > >> >> >> >>>> reference > >> >> >> >>>> potentials for CG. > >> >> >> >>>> > >> >> >> >>>> .tpr is binary, i.e. you cannot read it. > >> >> >> >>>> > >> >> >> >>>> csg_stat - I am not sure, better ask developers, but I would > >> >> >> >>>> expect > >> >> >> >>>> it > >> >> >> >>>> to come from the trajectory. > >> >> >> >>>> > >> >> >> >>>> V > >> >> >> >>>> > >> >> >> >>>> > >> >> >> >>>> > >> >> >> >>>> On 13 Aug 2013, at 10:56, [email protected] wrote: > >> >> >> >>>> > >> >> >> >>>> Sorry, my cell dimension was 95A (I was running temperature > >> >> >> >>>> annealing > >> >> >> >>>> so > >> >> >> >>>> the box dimension changed). Where in tpr file can I see the > >> >> >> >>>> actual > >> >> >> >>>> cell > >> >> >> >>>> dimension? I mean from where csg_stat takes the cell > dimension > >> >> >> >>>> to > >> >> >> >>>> calculate > >> >> >> >>>> RDFs? > >> >> >> >>>> > >> >> >> >>>> Steven > >> >> >> >>>> > >> >> >> >>>> W dniu wtorek, 13 sierpnia 2013 10:49:13 UTC+1 użytkownik > >> >> >> >>>> Valentina > >> >> >> >>>> Erastova napisał: > >> >> >> >>>>> > >> >> >> >>>>> Hi, > >> >> >> >>>>> > >> >> >> >>>>> I am a little confused there. > >> >> >> >>>>> > >> >> >> >>>>> You need to have a well equilibrated system before you CG. > My > >> >> >> >>>>> way > >> >> >> >>>>> is > >> >> >> >>>>> to > >> >> >> >>>>> run NPT, but make sure that the system is equilibrated and > V > >> >> >> >>>>> is > >> >> >> >>>>> also > >> >> >> >>>>> constant. > >> >> >> >>>>> > >> >> >> >>>>> When you start CG, I would use NVT, check what is the P > >> >> >> >>>>> (normally > >> >> >> >>>>> completely off), then do a P-correction. > >> >> >> >>>>> > >> >> >> >>>>> Good luck, V > >> >> >> >>>>> > >> >> >> >>>>> > >> >> >> >>>>> On 13 Aug 2013, at 10:37, <[email protected]> > >> >> >> >>>>> wrote: > >> >> >> >>>>> > >> >> >> >>>>> I think I know where is the reason... I set the wrong unit > >> >> >> >>>>> cell > >> >> >> >>>>> in > >> >> >> >>>>> my > >> >> >> >>>>> CG model. However, I run full atomistic in NPT and the > cubic > >> >> >> >>>>> unit > >> >> >> >>>>> cell was > >> >> >> >>>>> changing between 94A and 114A. Which shall I set up in my > CG > >> >> >> >>>>> run > >> >> >> >>>>> then which > >> >> >> >>>>> is in NVT? > >> >> >> >>>>> > >> >> >> >>>>> W dniu poniedziałek, 12 sierpnia 2013 17:02:36 UTC+1 > >> >> >> >>>>> użytkownik > >> >> >> >>>>> Valentina Erastova napisał: > >> >> >> >>>>>> > >> >> >> >>>>>> Hi, have you used scaling by by 1 /4 \pi r^2for bond and > by > >> >> >> >>>>>> 1/ > >> >> >> >>>>>> sin > >> >> >> >>>>>> (\theta) for angle distribution? > >> >> >> >>>>>> > >> >> >> >>>>>> I assume those are angles & bonds. > >> >> >> >>>>>> > >> >> >> >>>>>> Best, > >> >> >> >>>>>> V > >> >> >> >>>>>> > >> >> >> >>>>>> > >> >> >> >>>>>> > >> >> >> >>>>>> On 12 Aug 2013, at 16:39, [email protected] wrote: > >> >> >> >>>>>> > >> >> >> >>>>>> Dear Users, > >> >> >> >>>>>> > >> >> >> >>>>>> I run 7500 iterations of the protein with 5 different bead > >> >> >> >>>>>> types > >> >> >> >>>>>> so > >> >> >> >>>>>> 15 > >> >> >> >>>>>> distributions. Please, see attached 4 examples - black > >> >> >> >>>>>> atomistic, > >> >> >> >>>>>> red > >> >> >> >>>>>> coarse-grained. > >> >> >> >>>>>> 14 of them have exactly the same shapes as atomistic but > have > >> >> >> >>>>>> higher > >> >> >> >>>>>> values and they stopped converging after 6000 steps. Only > one > >> >> >> >>>>>> (attached) > >> >> >> >>>>>> perfectly converged. > >> >> >> >>>>>> May that be the reason of the one converged so its a wrong > >> >> >> >>>>>> distribution so the other cannot? Or everything is ok and > I > >> >> >> >>>>>> just > >> >> >> >>>>>> to > >> >> >> >>>>>> normalize them? > >> >> >> >>>>>> > >> >> >> >>>>>> Thank you, > >> >> >> >>>>>> > >> >> >> >>>>>> 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. > >> >> >> >>>>>> For more options, visit > >> >> >> >>>>>> https://groups.google.com/groups/opt_out. > >> >> >> >>>>>> > >> >> >> >>>>>> > >> >> >> >>>>>> <Dsitributions_RDF.png> > >> >> >> >>>>>> > >> >> >> >>>>>> > >> >> >> >>>>> > >> >> >> >>>>> -- > >> >> >> >>>>> 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. > >> >> >> >>>>> For more options, visit > >> >> >> >>>>> https://groups.google.com/groups/opt_out. > >> >> >> >>>>> > >> >> >> >>>>> > >> >> >> >>>>> > >> >> >> >>>>> > >> >> >> >>>> > >> >> >> >>>> -- > >> >> >> >>>> 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. > >> >> >> >>>> For more options, visit > >> >> >> >>>> https://groups.google.com/groups/opt_out. > >> >> >> >>>> > >> >> >> >>>> > >> >> >> >>>> > >> >> >> >>>> > >> >> >> >>> > >> >> >> >>> -- > >> >> >> >>> 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. > >> >> >> >>> For more options, visit > >> >> >> >>> https://groups.google.com/groups/opt_out. > >> >> >> >>> > >> >> >> >>> > >> >> >> >>> > >> >> >> >>> > >> >> >> >> > >> >> >> >> -- > >> >> >> >> 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. > >> >> >> >> For more options, visit > https://groups.google.com/groups/opt_out. > >> >> >> >> > >> >> >> >> > >> >> >> >> > >> >> >> >> > >> >> >> > -- > >> >> >> > 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. > >> >> >> > 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. > >> >> > 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. > >> > 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. > > 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. For more options, visit https://groups.google.com/groups/opt_out.
