[phenixbb] phenix.elbow

Nigel W. Moriarty nwmoriarty at lbl.gov
Mon Jun 2 12:17:44 PDT 2008


Thanks for your patience while I was away.  I have answered your 
questions below.
> When you set the --xyz option along with --clean then your .xyz file 
> is deleted. Of course the easy solution is simply to not use --clean 
> but I imagine that it is a minor bug since it doesn't happen with 
> --tripos.
Sure, I put this in to quickly delete so files.  I think its better to 
use --no-output which doesn't write most of the files to disk.  I'll 
look into --clean and may remove it.
> Secondly I have a slightly larger problem. When an isoSMILES with more 
> than one molecule is inputted then phenix.elbow joins them up to a 
> single molecule (or places them so closely that all programs, 
> including phenix.elbow, interprets them as one molecule when 
> converting to a format such as mol2 with explicit bonds).
> For an example try
> phenix.elbow --output=009_3085017 --write_hydrogens=false --overwrite 
> --clean --tripos --smiles='C=CCN.C1C(O1)CCl'
> The period is supposed to designate two seperate molecules.
You are correct that a period means two separate molecules.  If you are 
using eLBOW to generate restraints for a protein refinement, you should 
never pass two molecules (via SMILES) because of the nature of the 
lookup that is required for a ligand and its restraints.  That is, a 
single molecule such have a single three character code.  Also, the 
quantum code does not rely on explicit bond definitions to calculate the 
molecule energy.

Having said that, eLBOW does operate on a PDB with multiple ligands in a 
sensible fashion so if you can provide a case for running

phenix.elbow  --smiles="C=CCN.C1C(O1)CCl"

rather than running

phenix.elbow  --smiles="C=CCN"
phenix.elbow  --smiles="C1C(O1)CCl"

and some sort of rationale I'm sure I can support it.


Nigel W. Moriarty
Building 64R0246B, Physical Biosciences Division
Lawrence Berkeley National Laboratory
Berkeley, CA 94720-8235
Phone : 510-486-5709
Fax   : 510-486-5909
Email : NWMoriarty at LBL.gov
Web   : CCI.LBL.gov

More information about the phenixbb mailing list