The Silicon Graphics Teaching Laboratory
has been replaced by the
Chemical Information Laboratory
and this information is for historical interest only
Practical Six: Databases
(1) Searching the Cambridge Crystallographic Database
(a) Type 'quest filename' to start the search program, which is called
Quest. Draw this molecule as the substructure which the program will
search for in the database.
(b) After drawing the substructure, click DEFINE-STRUCTURE, to make
the program remember it, then click on TO-SEARCH. You now tell the
program what you want it to remember of the structures that it finds. Click
on FDAT and FCON, to remember to geometry and the connectivity of the
structures.
(c) Click on QUEST. We are only interested in structures whose three
dimensional structure is in the database, so click on INSIST-ON-
COORDS. Click on the picture of the substructure that should be visible, to
define the search question. Now click on START-SEARCH.
Be careful. It is not possible to interrupt a search once it has begun,
unless you find a structure. It is best, therefore, to begin with a loosely
defined search, which is likely to find many structures, rather than one
which may be too tightly defined. If you discover that you are finding too
many structures, you can easily stop and redefine your search. If your
search criterion is so tight that you find no structures, you will have to wait
until the program has looked through the entire database, before you can
try again.
(d) The program will search through the entire database to find structures
matching the substructure you have specified. Watch the original window
to see how it is progressing. When a structure is found, you can rotate it to
check it is what you want. Then click on KEEP, to remember the data.
(e) When the entire database has been searched, exit from Quest. We can
look at the structures that have been found by turning them into
MacroModel files. Type 'csd2mmod filename' to do this. There will now be
a file called filename.mmod which contains all the structures that were
found during the search.
(f) Read these structures into MacroModel, and calculate the average
oxygen-boron bond length. How strong is the association? Molecular
orbital calculations suggest that fluorine or other electronegative groups on
the boron will tend to eclipse the carbonyl. Is the data you have collected
consistent with this?
The Brookhaven Protein database is stored in the directory
'/usr/bigdisk/brookhaven'
Type 'ls
/usr/brookhaven' to get a list of the files and their cryptic code names.
These files can be read by MacroModel.
Copy the file to your own directory, type "gunzip filename.pdb.gz", and then read it using MacroModel, after selecting the PDB option in the Read window.
Alternatively, it can be searched using quest.
Type "quest filename -db PDB". When the search is complete,
type "pdbget -l -d -f filename.pid" to copy the
pdb files that have been found into your current directory.
A small selection, including index files, from the Brookhaven database
is in /usr/local/brookhaven directory. It will be quicker to look in this
directory than in the /usr/brookhaven directory.
| 
