VMD-L Mailing List

From: Peter Freddolino (petefred_at_ks.uiuc.edu)
Date: Mon Aug 04 2008 - 11:50:31 CDT

Hi Alan,
I'm out of town for the week and thus may not have time to look at these
in much detail until I get back. However, I would note the following
likely explanation:
1-4scaling only affects 1-4 interactions; if you're calculating an
*interaction* energy rather than a full system energy, only the pairwise
interactions between atoms in group A and those in group B will be
calculated. Since none of these are 1-4 interactions (unless you have a
covalently attached ligand) one would not expect any influence from
1-4scaling. Have you ever not seen a difference where you expect one in
a case where 1-4 interactions are actually being calculated?

Best,
Peter

Alan wrote:
> Hi Peter!
>
> So I did a test here and much more. I attached here a zip with a
> script and input files case you or someone else want to reproduce my
> results.
>
> To summarise. I got the results you mentioned here only if I set the
> whole system (protein, ligand, ions and water). If I try only vdw and
> elec for interactions between protein and ligand, then I got NO
> differences, be '1-4scaling' 1.0 or 0.8333. Can it be because of a so
> tiny difference since just few atoms are involved? I doubt. I
> put '1-4scaling' 0.1 or 10.0 and saw no difference.
>
> Besides I did another test. I ran 2 simulations of my system: one
> with '1-4scaling' 1.0 and other 0.8333. Log files shows differences as
> expected. Then I used NAMDenergy to recalculate energies for every
> frame of my dcd file for whole system, expecting to reproduce the same
> values (or very close to) in log simulations.
>
> Table Energy Results Last Step
> ETITLE: TS BOND ANGLE
> DIHED IMPRP ELECT VDW
> run_namd_amb: ENERGY: 1200 457.5011 1266.7038
> 2086.4423 0.0000 -88731.3521 10429.6362
> run_namd_def: ENERGY: 1200 518.8044 1347.3119
> 2096.4653 0.0000 -87295.0891 10529.5426
> namdEner_amb: ENERGY: 1200 457.5016 1266.7037
> 2086.4425 0.0000 -88730.9256 10429.6375
> namdEner_def: ENERGY: 1200 457.5016 1266.7037
> 2086.4425 0.0000 -87080.9837 10429.6375
>
> where:
> - 'run_namd' is simulation did with NAMD, 'namdEner' is energy
> recalculation did via NAMD with conf from NAMDenergy.
> - '_amb' is for '1-4scaling' 0.83333 and '_def' is
> for '1-4scaling' 1.0 (default)
> Note that I try the best I could to reproduce the exact same
> parameters in namdenergy conf file that I used to run the simulations
> (all can be verified in attachment, it would be rather a relief is I
> did a mistake there).
>
> So, looking at table above one can see that 'namdEner_amb' reproduces
> 'run_namd_amb' but I cannot say the same for 'namdEner_def'
> x 'run_namd_def'.
>
> I mean, I don't believe that NAMD has something suspect about it but I
> would love to understand why those results.
>
> Many thanks in advance.
>
> Alan
>
> On Fri, Aug 1, 2008 at 2:24 PM, Peter Freddolino <petefred_at_ks.uiuc.edu
> <mailto:petefred_at_ks.uiuc.edu>> wrote:
>
> Hi Alan,
> I tried a test similar to what you're doing, but I get the
> expected difference in electrostatic energy with and without
> 1-4scaling set:
>
> petefred_at_algiers ~/vmdwork/plugins/namdenergy =) $ namd2
> namd-temp.namd | grep ^ENERGY:
> ENERGY: 0 126.2869 269.3353 419.3776
> 0.0000 -23272.6906 2774.5828 0.0000
> 0.0000 0.0000 -19683.1080 0.0000
> -19325.0163 -19325.0163 0.0000
> petefred_at_algiers ~/vmdwork/plugins/namdenergy =) $ namd2
> namd-temp-noamb.namd | grep ^ENERGY:
> ENERGY: 0 126.2869 269.3353 419.3776
> 0.0000 -22741.4070 2774.5828 0.0000
> 0.0000 0.0000 -19151.8244 0.0000
> -18793.6619 -18793.6619 0.0000
> petefred_at_algiers ~/vmdwork/plugins/namdenergy =) $ diff
> namd-temp.namd namd-temp-noamb.namd
> 16,17c16,17
> < 1-4scaling 0.833333
> < scnb 2
> ---
> > #1-4scaling 0.833333
> > #scnb 2
> petefred_at_algiers ~/vmdwork/plugins/namdenergy =( $
>
>
> If you just test a pair of config files with/without this setting
> outside of namdenergy, do things work ok for you?
>
> Best,
> Peter
>
>
> Alan wrote:
>
> Thanks Peter.
>
> Pretty odd I know.
>
> And yes, I am sure my modified namdenergy.tcl is working
> because it
> gives error if put a wrong instruction. Besides, if you see
> the log
> files, you can clearly see 1-4scaling 0.8333 in one and 1.0 in the
> other. And to be completely sure, I was restarting VMD between
> tests,
> so it would source the modified file.
>
> I will test you code. But I am starting to think it can be
> NAMD the
> culprit. Thanks a lot.
>
> Cheers,
> Alan
>
> On Fri, Aug 1, 2008 at 5:59 AM, Peter Freddolino
> <petefred_at_ks.uiuc.edu <mailto:petefred_at_ks.uiuc.edu>> wrote:
>
>
> Hi Alan,
> odd... this should certainly work. This sounds silly, but
> are you sure
> you're modifying the copy of namdenergy.tcl that's part of
> the installation
> you're using? If so, could you try sourcing this tcl file
> prior to opening
> the namdenergy gui in a new vmd session?
> I'm attaching the modified version of namdenergy which is
> now in cvs, which
> should hopefully be more convenient for all users.
>
> Please let me know if sourcing the modified script still
> doesn't help.
>
> Peter
>
> Alan wrote:
>
>
> Hi Peter,
>
> Thanks for your reply.
>
> First I ran namdenergy GUI. Then I modified
> vmd/plugins/noarch/tcl/namdenergy1.2/namdenergy.tcl
> and included:
>
> puts $namdconf "exclude\t\t\t scaled1-4"
> puts $namdconf "1-4scaling\t\t\t 0.833333" <---- added
> puts $namdconf "outputname\t\t ${jobname}-temp"
>
> Then, open vmd, ran namdenergy GUI with same set up
> (putting 'namd2'
> in 'Temp File Prefix' and 'Output File' 'abc2.txt').
>
> I am attaching a zip file with files generated by
> debug option.
>
> Summarising, although log files are different about
> '1-4scaling', the
> rest and abc.txt and abc2.txt are identical.
>
> I am using latest version of Namd (2.6) and VMD
> (1.8.6) for Mac intel.
>
> I hope you can tell me what's going on here.
>
> Many thanks in advance.
> Alan
>
>
> On Thu, Jul 31, 2008 at 7:15 PM, Peter Freddolino
> <petefred_at_ks.uiuc.edu <mailto:petefred_at_ks.uiuc.edu>>
> wrote:
>
>
>
> Hi Alan,
> could you describe how you used namdenergy after
> you ran with the debug
> option and modified namd-temp.namd? Did you try to
> run it through the gui
> again or do it externally? I'm trying to verify
> that you didn't do
> anything
> to cause it to *re*write the .conf file prior to
> running namd.
> BTW, I've modified the version of namdenergy in
> cvs to automatically set
> 1-4scaling and sncb appropriately if you're using
> amber parameters, which
> should make life easier.
> Best,
> Peter
>
> Alan wrote:
>
>
>
> Dears,
>
> I use NAMD for carrying on a simulation with
> amber99 ff.
>
> The I open it in VMD:
>
> vmd -parm7 ComplexAmber.prmtop -rst7
> ComplexAmber.inpcrd run_namd.dcd
>
> and call namdenergy (dry-run) with debug
> option so I can modify
> namd-temp.namd file.
>
> What I don't understand is why if I add and
> set 1-4scaling to 0.5,
> 0.8333, 1 or whatever, the plot is exactly the
> same. I mean, if I run
> a simulation with NAMD using amber input files
> I know that I have to
> set 1-4scaling to 0.8333, but with namdenergy
> it seems to be ignored.
>
> I know that namdenergy will call NAMD to
> calculate energies all again
> for a given simulation (among other reasons to
> unscramble PME), but I
> don't realise why parameter "1-4scaling" is
> not taken in account. How
> do I know that my vdw energies are correct
> according to amber
> approach?
>
> Many thanks in advance.
>
> Alan
>
>
>
>
>
>
>
>
>
>
>
>
> --
> Alan Wilter S. da Silva, D.Sc. - CCPN Research Associate
> Department of Biochemistry, University of Cambridge.
> 80 Tennis Court Road, Cambridge CB2 1GA, UK.
>
>
> http://www.bio.cam.ac.uk/~awd28
> <http://www.bio.cam.ac.uk/%7Eawd28><<
>
>
>
>
>
> --
> Alan Wilter S. da Silva, D.Sc. - CCPN Research Associate
> Department of Biochemistry, University of Cambridge.
> 80 Tennis Court Road, Cambridge CB2 1GA, UK.
> >>http://www.bio.cam.ac.uk/~awd28 <http://www.bio.cam.ac.uk/%7Eawd28><<