Docking and density

[Rick Venable]

On 1 Aug 1995 19:17:01 GMT Jeffrey S. Taylor posted:
> Recently I've spent some time building a model of two proteins 'docked'
> together and now I'm spending more time trying to convince myself and
> my Professor that the model is reasonable.  An idea my Prof' had is to
> examine the density of the protein and compare it to the density at the
> interface of the two.  In a properlly docked model he reasons, the
> density should be constant, that is to say, the interface between the 
> two should have the same packing density as the protein itself.  

Possibly, but consider how to handle water molecules carefully.

> 	Does anyone know of a program to calculate density from atom
> coordinates that will help me with this?  Any other ideas for checking
> the model's validity would be appreciated.  I should say that I built
> the model with CHarMM, so the energetics are basically good, with 
> attention payed to charge and hydrophobic complementarity.

I've calculated densities with CHARMM many times; there's not a formal
command to do so, but all the tools are present.  The 3 key features
required are (1) use of ?XXXX substitutions to access CHARMM internal
variables, (2) use of user defined variables, e.g. @9, and (3) use of
atom SELEction, especially the obscure 'PROP 1' token.  I could provide
some example scripts to anyone who wants more detail, but here's a quick
example to calculate the density a 5 A cube centered at the origin:

! DEFINE PRINCIPLE CARTESIAN SLABS WITH PROP 1 ( == WMAIN )
scalar wmain = x
define xcut sele ( prop 1 .ge. -2.5 ) .and. ( prop 1 .le. 2.5 ) end
scalar wmain = y
define ycut sele ( prop 1 .ge. -2.5 ) .and. ( prop 1 .le. 2.5 ) end
scalar wmain = z
define zcut sele ( prop 1 .ge. -2.5 ) .and. ( prop 1 .le. 2.5 ) end
! GET THE MASSES, AND THE TOTAL THE ATOMS IN ALL 3 SLABS (A BOX)
scalar wmain = mass
scalar wmain stat sele xcut .and. ycut .and. zcut end
! PUT THE TOTAL IN USER VAR 9; DIVIDE BY VOL IN A**3; CONVERT TO G/CC
set 9 ?STOT
divi 9 by 125.
mult 9 by 1.66065

Note that 'scalar wmain = mass' must be replaced with a series of commands
to set the atomic numbers into WMAIN in order to get electron densities:

scalar wmain set 6.0 sele type C* end        [ etc. ]

as well as omitting the final multiplication.

--
Rick Venable                  =====\     |=|    "Eschew Obfuscation"
FDA/CBER Biophysics Lab       |____/     |=|
Bethesda, MD  U.S.A.          |   \  /   |=|  / Not an official statement \
rvenable at deimos.cber.nih.gov       \/    |=|  \  or position of the FDA.  /