From: Jérôme Hénin (jhenin_at_ifr88.cnrs-mrs.fr)
Date: Wed Jul 11 2012 - 10:20:58 CDT
Ipsita,
First, your variable involves several thousand atoms. Do you really
need so many? This could slow down the calculation dramatically. Try
to express the same quantity using a smaller number of atoms (e.g.
picking just one atom per molecule or per residue).
Your input describes a form of steered MD calculation, and an
extremely fast one: the restraint will move all the way in only 400 MD
timesteps. Maybe you intended to do slower steering, or maybe you
wanted to split the transition into 400 stages? See targetNumStages in
the documentation for details.
Your variable describes an isotropic distance. Did you want to look at
a z projection instead? If yes, see the distanceZ component in the
doc.
Finally, unlike in "traditional SMD", you don't need position
restraints because the variable is an internal coordinate of the
system. It is independent of global translations.
Best,
Jerome
On 11 July 2012 13:48, ipsita basu <ibasu788_at_gmail.com> wrote:
> Sir,
> I want to umbrella sampling in my membrane protein simulation to
> calculate potential mean force. My colvar configuration file is:
> colvarsTrajFrequency 200
> colvar {
> name length
> distance {
> group1 {
> atomNumbersRange 94313-98031
> }
> group2 {
> atomNumbers 1503 2307 2709
> }
> }
> }
> harmonic {
> name lenpot
> colvars length
> centers 17.0
> forceConstant 1.0
> targetCenters 80.0
> targetNumSteps 400
> }
>
> I want to vary the z-distance between group 1 and group 2.
> I put the initial zav value of group 1 in centers (centers 17.0)
> and the zav value of group in targetCenters.
> Now I am just a beginner in umbrella sampling and I've some queries
> about this. My questions are:
> 1. Is that right what I'm doing or I misunderstood?
> 2. What should be the value of targetNumsteps? Is it should be the
> value during which the distance will reduce by the given width?
> 3. should the rest system will be position restrained?
>
> Any type of help will be really appreciated.
> If the questions are bizarre then I'm sorry.
>
> Thanking you,
> Yours faithfully
> Ipsita
> --
> Ipsita Basu
> Research Fellow
> c/o : Dr. Chaitali Mukhopadhyay
> Rajabazar Science College
> 92 APC Road
> Kolkata - 700009
>
This archive was generated by hypermail 2.1.6 : Tue Dec 31 2013 - 23:22:14 CST