[buster-discuss] how to handle user defined stereochemical dictionaries

[alebus at pasteur.edu.uy]

Dear BUSTER community,

I am having some problems refining a protein that contains FAD.
Some of the 'available' dictionaries (such as ccp4) seem to have some  
errors (adenine N6 is not included in the corresponding planarity  
restraint, etc)
in the end, I'm obtaining some distorted geometries (res 2.8Å, SG 18)

So, two questions:

1. What is the proper way to use a user-defined dictionary?

I tried including the -l option for the autobuster refine command,  
immediately followed by the name of the dictionary (both a .cif,  
refmac style; or one .dic derived from that one using MakeTNT).
Autobuster refine will tell me in the beginning "Dictionary versions  
picked up from NOTE BUSTER_RESTRAINT_DICT cards as: ... ------  
restraints FAD by MakeTNT_2.4.0 from REFMAC dictionary...". (although  
I actually do not find the name of my dictionary anywhere...probably  
not reading it I guess)
But then, a few lines lower it will invariably say "Have no dictionary  
entry giving ccp4 atom types for residue types:  FAD SO4
        will use default ccp4 atom types CUNK  NUNK OUNK PUNK SUNK  
.... where possible."
(by the way, SO4 is actually in the 'regular' dictionaries: why can't  
it find it? I saw this question from Miguel Ortiz, but then no follow  
up on that...)

2. What is the procedure to look at this "fast"?
I mean, without actually running a full refinement protocol, just look  
at if buster actually read and kept the desired dictionary, and to  
which idealized model those restraints would actually be pointing to  
in further buster refinement cycles.

Thanks a lot for any advice!

Alejandro



-- 
Alejandro Buschiazzo, PhD
Research Scientist
Unit of Protein Crystallography
Pasteur Institute of Montevideo
Mataojo 2020
Montevideo 11400
URUGUAY

Phone: +5982 5220910 int. 120
Fax:   +5982 5224185