Re: [AMBER] aMD used for simulated small peptide with non-standard residues

From: Gabriel Jara <>
Date: Fri, 28 Oct 2016 15:16:31 -0200

>> I am interested in performing a conformational sampling of the peptide in
>> order to know its preferential conformations. I think about use
>> molecular dynamics (aMD) for that.

>Have you carried out ordinary MD simulations first? That might help
>the source of the problem. Also, which MD program are you using?

Hi David,

I have equilibrated the system first, then I ran a 20ns of MD in cpu and
then 1000ns (because it was my first attempt for sampling the system). The
problem I reported before was using problem PMEMD.cuda in amber16.

I delayed in answering because I have performed other simulations for
trying to be more specific.

I ran a aMD simulation in CPU, and I had the same problem. However, when I
put tol=0.000001, it seems corrected the crashing in simulation. However,
it still has a strong variation in temperature. I ran for 30ns. When I run
a aMD with GPU using the same input, after 9 ns i had the same problem.

Keeping a shake issue in mind, I also ran a aMD simulation in GPU without
shake and with a time step 1fs. I have 91ns of this simulation without
error and the temperature oscillate 3K.

For remembering, my system is a tetrapeptide with non-standard residues and
two protecting groups in the N and C terminal.It was parametrized using
R.E.D Serv. Development and the solvent is chloroform.

I have no idea what it is happening. Thank you for your helps and

Do you have any idea what could be wrong? maybe the parametrization i made
is not so good.



Dr. Gabriel E. Jara
Instituto de Química / UNICAMP
Rua Josué de Castro s/n
Cidade Universitária "Zeferino Vaz", Barão Geraldo
13083-861 Campinas, São Paulo, Brasil
AMBER mailing list
Received on Fri Oct 28 2016 - 10:30:02 PDT
Custom Search