Re: protein on solid surface

From: Giovanni Bellesia (giovanni.bellesia_at_gmail.com)
Date: Fri Apr 08 2011 - 13:45:31 CDT

Giacomo,

thanks for the prompt feedback. I didn't think about the boundary potential
although I used it in the past (in other simulations) via tclBC.
I am not sure I understood what you suggest in terms of distanceZ (surely
because I haven't been using the colvar module regularly in my
calculations).

Let's say that I want explore a small rectangular area within my surface.
and also that the area is defined by the positions of 4 atoms (A,B,C,D ...
those will be the 4 vertexes of the rectangle).

Something like this:

A---------------B
| |
| |
| |
| |
| |
C---------------D

-------------------> y-axis
|
|
|
v
x-axis

So, one group of atoms will certainly be the one defining the protein COM.
Since I would like to have a free energy map I have to define two collective
variables both involving in some way
the distance between the protein com and some point on the surface, right ?
At this point I am stuck ...
What I am thinking is to define:

(1) a distanceZ cvar considering the x-axis (see scheme above) and one atom
or group of atoms nearby vertex A.
     Set as lowerboundary the distance between the com of the protein and
the atom in A when the protein is "sitting" on A (let's call this LB)
     Set as upperboundary simply using Pitagora's as follows : sqrt(LB^2 +
(A-C)^2)

(2) a distanceZ cvar considering the y-axis and once again A with same
lowerboundary and with upperboundary = sqrt(LB^2 + (A-B)^2)

Does this make any sense to you ?
How is this different from what you were suggesting ?

Thanks

Giovanni

On Fri, Apr 8, 2011 at 10:21 AM, Giacomo Fiorin <giacomo.fiorin_at_gmail.com>wrote:

> Hello Giovanni, I think Cartesian coordinates of the protein's center of
> mass can be used as collective variables, and should be also very intuitive
> to analyze. Use a distanceZ-based colvar in all three cases, redefining the
> "axis" each time. Along x and y (surface-parallel) you may want to define
> their period, to take advantage of the periodic boundary conditions. To
> prevent the protein from coming off, you can set a boundary potential along
> z, which doesn't kick in within a tolerance distance of your choice: this
> way the protein can be free to jump over the corrugations of the surface,
> but not come off it permanently.
>
> Giacomo
>
> ---- ----
> Giacomo Fiorin - Postdoctoral Researcher
> ICMS - Institute for Computational Molecular Science - Temple University
> 1900 N 12 th Street, Philadelphia, PA 19122
> giacomo.fiorin_at_temple.edu
> ---- ----
>
>
>
>
> On Fri, Apr 8, 2011 at 10:48 AM, Giovanni Bellesia <
> giovanni.bellesia_at_gmail.com> wrote:
>
>> Hi all,
>>
>> I have a small protein "sitting" on a solid surface made up by cellulose
>> chains and I would like
>> my protein to "explore" a small portion of the surface itself.
>>
>> The idea is use ABF (and/or metadynamics) to generate a
>> 2D free energy map to find out which "spots" on the cellulose surface
>> correspond to free energy minima.
>> I was wondering if anyone has suggestions in terms of choice of a pair of
>> suitable collective variables.
>> In more detail, I was wondering how to choose a pair of cvars in a way
>> that the ABF (or metadynamics) calculation drives the protein
>> to "move" on the surface without "taking off" and moving back in the bulk
>> water topping the surface.
>>
>> The protein surface interaction is hydrophobic in nature and in my
>> starting configuration
>> the hydrophobic groups in my protein are already in contact with the
>> surface.
>> All this to say that I don't need to explore possible alternative
>> protein-surface contacts.
>> I just need the protein to simply "slide" on the surface with the same
>> relative orientation (i.e. hydrophobic groups in contact with the surface).
>>
>> Thanks for your help
>>
>> Giovanni
>>
>
>

This archive was generated by hypermail 2.1.6 : Mon Dec 31 2012 - 23:20:06 CST