[AMBER] AMBER Alanine tutorial (different outputs from AMBER15 and AMBER18)

From: <leifepeterson.sbcglobal.net>
Date: Sun, 16 Sep 2018 14:30:24 -0500

Thanks for the feedback. Something major I noticed for the Alanine Tutorial
run with AMBER18Tools is that:

1. after replacing the ntb= commands back into the two files and going
through Leap and Sander runs using commands that are identical to the
tutorial, once the Production MD run is completed, the size of trajectory
file 03_Prod.mdcrd was only 6.9MB, whereas the file size form the tutorial
is 13.9MB. When the larger file from the tutorial page (13.9MB) is input
into VMD after the prmtop file is input, the movie works, however, using my
smaller 03_Prod.mdcrd (6.9MB) file, there is an error on input into VMD.

2. When plotting the summary.DENSITY file, there is simply a straight line
with constant slope, so this would imply nothing is being done
computationally regarding an increasing box size.
 
Since the online Alanine tutorial is based on AMBER15, my guess is that
program modifications preclude one from obtaining the same tutorial results
using AMBER18.

Have any AMBER developers rerun that tutorial using AMBER18?

-----Original Message-----
From: Matias Machado <mmachado.pasteur.edu.uy>
Sent: Sunday, September 16, 2018 10:35 AM
To: AMBER Mailing List <amber.ambermd.org>
Subject: Re: [AMBER] RESOLVED: peek_ewald_inpcrd: Box info not found in
inpcrd

Dear ???,

If there is no box information in your coordinate file, then you have set
something wrong during the topology and coordinates generation...

Did you visually inspect the system before running? is the water present?
Because the command "solvatebox foo TIP3PBOX 10.0" should add the box
information into the system (at the last line in the inpcrd file)...

I'm not an AMBER developer to tell you about the code check points, but
without box information (the usual set up for igb>0) you can't apply any pbc
condition, so by removing "ntb" you are setting no periodicity (ntb=0, which
is the default in igb>0)...

So, removing "ntb" may not be a way to solve the problem but a way to avoid
an error message, which indeed means hiding the real problem...

I think this example is a very good reason to keep the ntb flag, so you can
really check on what you are doing.

Hence, I think the following statement in the manual may not be complete
true:

"There is no longer any need to set this variable, since it can be
determined from igb and ntp parameters."

Because if you mess up with the box information then you may end up running
at unwanted conditions without knowing...

Best,

Matias

------------------------------------
PhD.
Researcher at Biomolecular Simulations Lab.
Institut Pasteur de Montevideo | Uruguay
[http://pasteur.uy/en/laboratorios-eng/lsbm]
[http://www.sirahff.com]

----- Mensaje original -----
De: leifepeterson.sbcglobal.net
Para: "AMBER Mailing List" <amber.ambermd.org>
Enviados: Domingo, 16 de Septiembre 2018 0:24:35
Asunto: [AMBER] RESOLVED: peek_ewald_inpcrd: Box info not found in inpcrd

This error was resolved for sander runs by simply removing the line with
"ntb=1," in the file 02_Heat.in and the line with "ntb=2," in the file
03_Prod.in

 

The User's Manual states that the ntb setting is not really needed anymore,
since other parameters can be used to obtain the information.

 

However, why doesn't a developer/curator remove the "ntb= ," lines from
these files in the Alanine Leap and MD tutorial.

_______________________________________________
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


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Sun Sep 16 2018 - 13:00:01 PDT
Custom Search