#21 a way to detect distance from surface

Bob Hanson

Rasmol has predefined selection parameter of amino
acids called surface and one called buried. Since so
much relies on interaction at surfaces it would be nice if
JMOL could select actual surface residues and buried

Another reason for such a request - I was looking at this
paper below and naively thought it might be easy to use
Chime to color/list residues based on atom depth. (It is
especially interesting since in general more conserved
residues usually tend to be solvent inaccessible.) I tried
to come up with a script that would start color coding
based on atom depth but realized the problem was that
after I had calculated the water accessible surface area
I couldn't figure out the command to select residues
within a certain distance from the calculated surface. I
looked at MDL's info on surface stuff but couldn't figure
it out.
Anyway it would be nice to think about for building into
Jmol since in the building process. A lot of drug design
and examination of interactions depends on surfaces.
Here is the paper:
Atom depth in protein structure and function
Alessandro Pintar, Oliviero Carugo and Sandor Pongor
Atom depth, originally defined as the distance between a
atom and the nearest water molecule surrounding a
protein, is a
simple but valuable geometrical descriptor of the protein


  • Michael

    Logged In: YES

    Unfortunately, I cannot get to this article.

    I need to understand the complexity of determining the
    distance from the surface. To me it sounds complicated.

  • molvisions

    Logged In: YES

    the Rasmol/Chime functionality described here is *quite* simplistic.
    buried and surface are just pre-defined sets of residues that are generally
    found in one locale or the other. it does not acutally calculate these
    properties for each structure file!

    that said, I do agree that powerful surface generation and rendering would
    be an excellent addition to Jmol. I believe GRASP has one of the best set
    of surface tools around - my opinion, of course, but a good place to poke
    around if possible.

  • Bob Hanson
    Bob Hanson

    • labels: --> Interface Improvements
    • assigned_to: nobody --> hansonr
  • Bob Hanson
    Bob Hanson

    Logged In: YES

    we could do this relatively easily now. It just has to do
    with how many points are present from a dots-based
    rendering of the solvent-accessible surface. Atoms/Groups
    could be colored based on distance from the surface or
    selected based on some specified distance from the

    Could be very interesting.

  • Bob Hanson
    Bob Hanson

    Logged In: YES

    An old, old anonymous request finally accomplished.

    load 1crn.pdb
    select surfacedistance=0
    color white

    load 1a31.pdb
    color surfacedistance

    calculate surface +1.2 #default, but
    calculate surface +1.5 #customizable

    This uses the dots/geosurface fast-surface generator to
    simply count the number of dots; pretty simple!

    The distance value is an approximation -- just the minimum
    distance from an atom to a surface ATOM. So this results in
    all surface atoms having a value of 0, and all other atoms
    having a value of > 0.

    Note that you can say:

    select not solvent
    set dotsSelectedOnly
    calculate surface +1.2

    and any solvent water will NOT be inside the boundary.


  • Bob Hanson
    Bob Hanson

    • status: open --> closed