View source for Huge surfaces

PyMOL can render '''very large''' (huge) surfaces of proteins, using a few tricks.  First off, you should know that because of the size, you do not get the level of accuracy that you would for a smaller molecule.  However, if you need to represent a molecule (or object) with a few million atoms, this may be your option of choice.

= Example =
The output from the example below is shown below.  The PDB 1AON has nearly 60,000 atoms.  This isn't the largest PDB, but it's a good example.

<gallery>
Image:large_surface.png
Image:large_surface_approx.png
</gallery>

== Strategy 1: Low resolution Gaussian Map ==
<source lang="python">
# load a whopping big PDB
fetch 1aon, struct, async=0

# create a color spectrum over the object
spectrum count, selection=struct

# === now create a pseudo-fcalc map (a 3D volumetric scalar field) ===
# set the B-factors nice and high for smoothness
set gaussian_b_floor, 40

# ~10 A map resolution
set gaussian_resolution, 10

# ~10 A map spacing with a 10 A surrounding buffer
#  (you may need to vary this)
map_new map, gaussian, 10, struct, 10

# create a surface from the map
isosurface surf, map, 1.0

# now color the map based on the underlying protein
ramp_new ramp, struct, [0,10], [atomic, atomic]

color ramp, surf

# hide the ramp
disable ramp
</source>

== Strategy 2: CA-only model with oversized radii ==
<source lang="python">
# load a whopping big PDB
fetch 1aon, struct, async=0

# reduce to CA atoms only
remove struct & not guide

# create a color spectrum over the object
spectrum count, selection=struct

# increase VDW radii to compensate volume of missing non-CA atoms
alter struct, vdw=4

# increase VDW radius of solvent
set solvent_radius, 4, struct

# show molecular surface
as surface
</source>

[[Category:Performance]]