[gmx-developers] help for integrating minimization algorithm

[Mark Abraham]

Hi,

On Mon, Dec 5, 2016 at 7:30 AM Wilhelm Bender <wilhelm.bender at gmx.de> wrote:

> Hello,
>
> i found an entry point to implement my minimizer. I basically replaced the
> search direction inside do_steep() in minimize.c. Now I encounter another
> problem:
>
> as soon as I translate a sequence to a pdb-file I get an alpha-helix as a
> start structure. When I minimize by using the integrator "steep" gromacs
> seems to penalize anything that wants to break the alpha-helix.


Sure. Those hydrogen bonds make for a pretty steep-walled energy minimum,
and usually the minimizer's job is to follow the gradient.


> I read here
> and there that this may be a setting inside gromacs concerning restraints
> and constraints. The minimizer that I apply is a (quasi) global optimizer,
> therefore I need to be able to find a configuration of my input files for
> mdrun such that I can aggressively alter the starting structure without
> getting problems with any restraints and constraints ripping apart.
>
> Do you have a suggestion for a setup in order to do global optimization?
>

Well, you could avoid using constraints, and fail to add any restraints,
and otherwise choose emstep to be small.

I presume you are operating in implicit/no solvent in order that you can
get anything done - there's a reason why good MC move sets for compact
condensed-phase models are elusive.

Mark

Best Regards and Thank you for any help,
>
> Wilhelm
>
> -----Ursprüngliche Nachricht-----
> Von: gromacs.org_gmx-developers-bounces at maillist.sys.kth.se
> [mailto:gromacs.org_gmx-developers-bounces at maillist.sys.kth.se] Im Auftrag
> von David van der Spoel
> Gesendet: Donnerstag, 21. Juli 2016 18:33
> An: gromacs.org_gmx-developers at maillist.sys.kth.se
> Betreff: Re: [gmx-developers] help for integrating minimization algorithm
>
> On 21/07/16 11:42, Wilhelm Bender wrote:
> > Hello,
> >
> >
> >
> > Im new to gromacs development therefore I need a few hints about how
> > to extend functionalities concerning the energy minimization. I
> > basically found my entry point for integrating a minimization
> > algorithm that uses a gradient much like do_steep(). Can I somehow
> > take the gradient computed and calculate a norm of it and the add the
> > search direction to the current coordinates? Maybe there is a thread that
> I overlooked?
> >
> >
> How about do_cg (conjugate gradients)?
>
> >
> > Help would be appreciated J
> >
> >
> >
>
>
> --
> David van der Spoel, Ph.D., Professor of Biology Dept. of Cell & Molec.
> Biol., Uppsala University.
> Box 596, 75124 Uppsala, Sweden. Phone:  +46184714205 <018-471%2042%2005>.
> spoel at xray.bmc.uu.se    http://folding.bmc.uu.se
> --
> Gromacs Developers mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers or
> send a mail to gmx-developers-request at gromacs.org.
>
> --
> Gromacs Developers mailing list
>
> * Please search the archive at
> http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List before
> posting!
>
> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
>
> * For (un)subscribe requests visit
> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers
> or send a mail to gmx-developers-request at gromacs.org.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-developers/attachments/20161209/05001f72/attachment-0001.html>