[Dock-fans] inv_a error

Scott Brozell sbrozell at rci.rutgers.edu
Mon Jan 23 23:50:41 PST 2012


Hi,

On Thu, Jan 19, 2012 at 10:50:15AM +0100, Jose R. Valverde wrote:
> 	I am running a dock job using Zou GB/SA scoring, and I keep getting
> a number of errors in each pose calculation, all of the kind
> 
> 	Pure ligand inv_a error: 46     -20.749
> 	Receptor on ligand inv_a error: 33      -24.7508
> 
> 	I suspect this means the calculation results will be wrong and 
> should not be used. Right? Or is this just an informative message that
> can be disregarded?

After reading section 2.2 of
http://dock.compbio.ucsf.edu/DOCK_6/dock6_manual.htm#LiuetalJPhysChemB2004
this error indicates that the sum in eqs 21 and 22 has grown so big
that the multiplicative inverse of the Born radius of the atom,
1/alpha_i in the paper, aka gb_lig_inv_a[i] in the code, will be
negative.  To avoid this unphysical situation, the code assigns 0.1
to gb_lig_inv_a[i].

You should examine atoms 46 of the ligand and 33 of the receptor.
Perhaps there is a simple explanation and a simple fix.

good luck,
scott



More information about the Dock-fans mailing list