Hi,
First, if I'm doing the math right I think you're specifying the wrong
time between frames to 'diffusion'. If each trajectory is 5 ns, and
there are 100000 frames, then frames were written every .05 ps, not
1.0 ps. Since you're taking every 10th frame, you'll want to specify
the time argument as 0.5. I'll assume you have a reason for wanting to
write out so frequently (although it seems *way* too frequent for
calculating diffusion from MSD in my opinion!). This won't solve your
issue, but will ensure units are correct in your results.
Beyond this, there are three things you can try.
1) Upgrade to cpptraj from GitHub or AmberTools 20. The Amber16 code
is about 5 years old at this point - lots of bugfixes and enhancements
over those years.
2) See if maybe the imaging code in 'diffusion' is not working well
for some reason. Try to use 'unwrap' before 'diffusion', and add the
'noimage' keyword to the latter, e.g.
strip :WAT,Na+
unwrap
diffusion out average MSDs1tr_without54_WA time 0.5 :1-53,55-67
3) Try the 'stfcdiffusion' command, which uses separate code written
by Hannes Loeffler. Maybe you're hitting an issue in 'diffusion' not
present in 'stfcdiffusion'.
Hope this helps,
-Dan
On Mon, Jan 25, 2021 at 3:38 AM Pau Mayorga Delgado
<Pau.MayorgaDelgado.student.oulu.fi> wrote:
>
> Dear amber group,
>
> Is there any issue with the diffusion command in cpptraj amber-16? For a 66-lipid micelle we get a plateau value that does not follow a linear relationship and we have not seen this in other systems. There seems not to be a drift in the electrostatic interactions. This is a NVT simulation.
>
> The commands in ptraj are as follows:
>
> trajin Simu1_run1.mdcrd 1 100000 10
> trajin Simu1_run2.mdcrd 1 100000 10
> trajin Simu1_run3.mdcrd 1 100000 10
> trajin Simu1_run4.mdcrd 1 100000 10
> trajin Simu1_run5.mdcrd 1 100000 10
> trajin Simu1_run6.mdcrd 1 100000 10
> trajin Simu1_run7.mdcrd 1 100000 10
> trajin Simu1_run8.mdcrd 1 100000 10
> strip :WAT
> strip :Na+
> diffusion :1-53,55-67 1.0 average MSDs1tr_without54_WA
>
> go
>
> In this case there are 8 coordinate files of 5ns each that amount to a 40ns MSD simulation. The MSD plot averaged over 4 simulations is presented below.
>
> [cid:bef89445-b394-4ba3-b66a-b72286421824]
>
> Thanks for the help.
>
> Best Regards,
>
> Pau
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Mon Jan 25 2021 - 06:30:04 PST
