Re: string method in NAMD 2.10

From: Thomas Evangelidis (
Date: Tue Sep 22 2015 - 05:29:16 CDT


I had the same query in past so I contacted Luca Maragliano. I am
forwarding below
​ communication
​, hopping​
that it will be useful for other people interested in the string method.

---------- Forwarded message ----------
From: Luca Maragliano <>
Date: 15 December 2014 at 15:43
Subject: RE: string method with swarms-of-trajectories - NAMD implementation
To: Thomas Evangelidis <>

Metadynamics and the string method have different purposes. in path cv
metadyna they (usually)
keep the initial path fixed and sample around it via US. If the initial
path is far from the actual MFEP
the space you need to map with US can be quite extended. With the string
method you relax the path,
and once you have it converged you can decide what you want to do with it.
The PMF calculation along
the path using mean forces is less expensive than an extensive 2D US

As another piece of info let me tell you about another code that has the
otf implemented

some of the documentation is in japanese but you can write Kei Moritsugu if
you have
questions, he will answer for sure.

For the implementation of the otf version, look at table 1 of our JCTC
paper (page 527).
The otf method is similar to the mean forces method except that steps 1-3
substituted by the concurrent evolution of the CVs and the physical
variables, coupled
via the harmonic potential. In other words, rather than doing step 1 with
fixed CV values
you evolve the CVs as well but on a different, slowed-down time-scale, i.e.
you give them
a high mass or friction. More details are in appendix B.
In terms of implementation, this amounts to modify the restraint dynamics
of step 1 by
introducing a line that evolves the CV values.

I hope this helps, but please feel free to ask more.



*From:* Thomas Evangelidis []
*Sent:* Monday, December 15, 2014 1:24 PM
*To:* Luca Maragliano
*Subject:* Re: string method with swarms-of-trajectories - NAMD

Dear Luca,

Thank you for the explanation. I still cannot find any info about the
implementation of String method in NAMD, neither in the publication you
cited nor in the NAMD bundle. Those applications of the method to membrane
proteins were conducted with NAMD or CHARMM?

I am more familiar with metadynamics, but since that requires a lot of
preparation and optimization of the path CV, I thought the String method
with swarms of trajectories would be more straightforward. Of course with
the latter you must do umbrella sampling at the end to get the PMF, whereas
with metadynamics you get it with no extra effort. If you feel comfortable
indicating the modification you mention on the swarms version, then I would
try the on-the-fly version, otherwise I would go with metadynamics.

thank you for your help,

On 15 December 2014 at 11:06, Luca Maragliano <>

Dear Thomas,

thank you for your interest. I think that the updates of NAMD concerning
the string method
were discussed in the paper here

and as far as I know they don't use the scripts attached to our JCTC (or at
least I never discussed
them with NAMD dev people).

There are now several applications of the string method to large membrane
proteins so yes I would
recommend to use it.

In our JCTC paper we show that the on-the-fly version (CPL 2007) is more
efficient than the other
two algorithms. I don't think that it is implemented in NAMD but it could
be easily done with
small modifications to the swarms version. Note also that the methods
calculate slightly different
paths (as discussed in the JCTC paper). Usually the differences are within
statistical fluctuations
but this cannot be said a priori for any application.

Please feel free to ask more if needed, I'd be happy to help where I can.



*From:* Thomas Evangelidis []
*Sent:* Monday, December 15, 2014 12:22 AM
*To:* Luca Maragliano
*Subject:* string method with swarms-of-trajectories - NAMD implementation

Dear Luca,

I working on a transporter and want to simulate the substrate-occluded->outward
facing conformational transition. In the release announcement of NAMD 2.10
it is stated that the String method with swarms-of-trajectories transition
pathways is supported. However, there is no official NAMD documentation
about this implementation. I guess they mean to use the scripts you provide
in the supplementary material of you recent publication, right?

In the paper you demonstrate the advantages with respect to the original
method, using the alanine dipeptide as a test system. Would you be confident
applying the same methodology on a system as large as a transmembrane

thanks in advance,

Thomas Evangelidis
Research Specialist
CEITEC - Central European Institute of Technology
Masaryk University
Kamenice 5/A35/1S081,
62500 Brno, Czech Republic


This archive was generated by hypermail 2.1.6 : Tue Dec 27 2016 - 23:21:20 CST