VMD-L Mailing List

From: Peter Freddolino (petefred_at_umich.edu)
Date: Mon Nov 09 2020 - 00:15:42 CST

I think one useful step would be to actually calculate the magnitudes of
the differences in coordinates between the original trajectory and the one
read/written by VMD. The energy differences that you've noted here are
tiny, on the order of less than one part per thousandth, and the energies
are going to be very sensitive to small changes in coordinates.
I also suspect that if you actually inspect the files before/after running
them through your procedure in vmd, you will find that they are in fact not
the same format. Since you're using netcdf files in amber, you have a
binary container format when you read them into VMD, but when you write
from VMD in crdbox format, you're converting to an ascii format with
different precision. Just because the file extension is .crd doesn't
actually mean that the format is the same; you can see this easily enough
by looking at them in any text viewer or editor. This seems likely to me to
be the issue. If you want to avoid any changes you'd be best off comparing
apples to apples, and having either netcdf-format or crdbox-format files
both before and after vmd sees the coordinates.
Best,
Peter

On Sun, Nov 8, 2020 at 10:56 PM Prathit Chatterjee <pc20apr_at_yahoo.co.in>
wrote:

> Dear Prof. Freddolino,
>
> Sorry for the delayed response.
>
> The precision of selected coordinate numbers to calculate the energy is
> the same in both the original AMBER coordinates and protein-only generated
> trajectory (in VMD), and the format is .crd, calculated with cpptraj,
> please see below:
>
> *This is for the original amber coordinates*:
> parm prot.top
> trajin /homes/epsilon/users/diem/project_t-ab/traj_1/production/crd/
> *T_310K_1_md001.crd* *1 10*
>
> #centering
> strip :WAT
> strip :Na+
> strip :Cl-
>
> center :1-43 mass origin
> image origin center familiar
>
> pairwise IE :1-43 out pairwise_IE.out cuteelec 0.0001 cutevdw 0.0001
> vmapout IE_vmap.out emapout IE_emap.out avgout IE_avg.out
> --------------
>
> *This is for the protein only generated coordinates*:
> parm prot.top
> trajin /homes/epsilon/users/pchatterjee/vmd_anal_prep/diem_1/
> *T_310K_prot_diem_1_100ns.crd* *1 10*
>
> pairwise IE :1-43 out pairwise_IE.out cuteelec 0.0001 cutevdw 0.0001
> vmapout IE_vmap.out emapout IE_emap.out avgout IE_avg.out
> -------------
>
> Also, *I generated the protein-only amber coordinates as follows in VMD
> (a prototypical example provided below)*:
>
> mol new
> /homes/epsilon/users/diem/project_t-ab/traj_10/production/whole.top type
> parm7
> mol addfile
> /homes/epsilon/users/diem/project_t-ab/traj_10/production/crd/T_310K_10_md$n.crd *type
> netcdf* waitfor all
>
> set frames [molinfo top get numframes]
> set nf [expr {[molinfo top get numframes]-1}]
>
> set a [atomselect top protein]
> animate *write crdbox* diem_10/PHF43_310K_prot_$f.crd beg 0 end
> $nf waitfor all sel $a
>
> I tried to write the protein only coordinates in several ways until
> reaching the above mentioned procedure, loading the original coordinates as
> netcdf, and writing the final protein only coordinates as crdbox.
> Likewise, the coordinates of the original and protein-only coordinates
> could be visualized properly without noticing any apparent unnatural bonds.
>
> Best Regards,
> Prathit
> On Sunday, 8 November, 2020, 01:35:42 pm GMT+9, Peter Freddolino <
> petefred_at_umich.edu> wrote:
>
>
> Dear Prathit,
> We really need more information on how you did these calculations. How
> were the energies calculated? And how exactly did you generate the
> trajectory in vmd? With such small differences it could be something as
> simple as a difference in the numerical precision at which coordinates were
> stored, depending on exactly what formats you used.
> Best,
> Peter
>
> On Sat, Nov 7, 2020 at 10:02 AM Prathit Chatterjee <
> pc20apr_at_remove_yahoo.co.in> wrote:
>
> Dear VMD experts,
>
> I had compiled a protein only amber trajectory of an explicitly solvated
> simulation with VMD, which I was able to visualize properly and run certain
> analyses with it in VMD as well.
>
> But, for cross-checking, I tried to calculate the pairwise interaction of the
> original AMBER coordinates, and the protein only generated trajectory
> (with VMD).
>
> The results, although very similar, are showing certain discrepancy as
> follows:
>
> [pchatterjee_at_node51 pairwise_IE]$ tail -f test*/pairwise_IE.out
> *==> test1/pairwise_IE.out <== [THIS IS THE ORIGINAL AMBER COORDINATE]*
> 1 -186.3642 -2973.0974
> 2 -189.9812 -2966.1259
> 3 -199.5370 -2957.7041
> 4 -196.3846 -2983.7931
> 5 -197.8264 -2981.1530
> 6 -184.0711 -2971.6430
> 7 -194.3659 -2970.1477
> 8 -189.0441 -2956.9308
> 9 -197.0864 -2948.2561
> 10 -196.4939 -3010.1367
>
> *==> test2/pairwise_IE.out <==[THIS IS THE PROTEIN ONLY GENERATED AMBER
> COORIDNATE IN VMD]*
> 1 -186.3943 -2972.9815
> 2 -189.9798 -2966.1573
> 3 -199.5455 -2957.6502
> 4 -196.3961 -2983.7881
> 5 -197.8009 -2981.1609
> 6 -184.0410 -2971.7176
> 7 -194.3681 -2970.0830
> 8 -189.0185 -2956.9181
> 9 -197.1040 -2948.2302
> 10 -196.4685 -3010.0947
>
> Although the global picture will be the same with such minor discrepancy,
> I just want to know for the sake of knowing what the underlying reason
> could be.
>
> Any comments/suggestions will be extremely helpful.
>
> Thank you in advance,
> Best Regards,
> Prathit
>
>