From: Jurgen F. Doreleijers (jurgenfd@gmail.com)
Date: Fri May 27 2005 - 10:07:46 CDT
Hi Gary,
Thanks for posting the message on the BMRB mailing list. Here's a howto that
works on Linux under tcsh shell:
-1- Install the programs tar, wget, and gawk if they aren't already.
-2- Get all the data in the FRED database in CNS format as they were used
for creating the RECOORD database. This is described below for NOE distance
restraints. Repeat it for the other types of NMR restraints that you want to
retrieve, e.g. dihedral angles.
# Get all converted NOE distance restraint lists in XPLOR/CNS format from
the FRED database in the NMR Restraints Grid.
* Select "filtered" for "Block detail".
* Select "-EMPTY-" for "File detail" (the first choice).
* Select show "Blocks".
* Enter 0 for "Program instances above".
* Do a submit.
* The resulting page shows a list of files containing all different data
types for/from many different programs. Select the link for the number of
blocks in the column labeled "XPLOR/CNS" on the row with for "Type, Subtype,
and Format": distance, NOE, ambi. The unambiguous and ambiguous NOE distance
restraints have been combined in FRED.
* The resulting page shows another table with the PDB entry codes (in
the column labeled "PDB_ID") of entries that have filtered NOE distance
restraints as part of the deposited data. To convert the table to a
comma-separated-values file, click the table icon just above the table and
save the file as "FRED_list_distance_NOE_ambi.csv" on your computer. Most
spreadsheet programs, e.g. Gnumeric and MS Excel, can read the file.
* To save the data, select the other icon. You will start to download a
zipped file that contains a separate file for each block in the selection
you made up to this point. The blocks are named only by their block
identifier (a unique number for each block in the database) and change over
time so please don't use them for future reference but use the saved csv
version of the table so you have additional info on the block as needed.
* Unzip the downloaded file into a target directory of choice on your
computer.
-3- Get the list of PDB entries in the FRED database from the file:
"FRED_list_distance_NOE_ambi.csv" you just saved. Select only the PDB codes
from this file into a new file by doing on Linux: gawk -F',' '{if (NR!=1)
print $3}' FRED_list_distance_NOE_ambi.csv > FRED_list.txt
-4- Get the refined models for all water rerefined structures from CNS in
RECOORD by looping over the entries: set list = ( `cat FRED_list.txt` ) echo
"Doing number of entries: $#list" foreach x ( $list )
echo "Getting ensemble for $x"
wget -q
http://www.ebi.ac.uk/msd-srv/docs/NMR/recoord/files/$x/coord_recalculated/cn
s/$x"_cns".tgz
end
-5- Uncompress to individual coordinate models as needed:
tar xvzf $x*_cns.tgz
Thanks again for your interest, cheers,
Jurgen
---
Jurgen F. Doreleijers, Ph.D.
BMRB/CESG, Univ. of Wisconsin-Madison, WI, USA
mailto:jurgenfd@gmail.com
You wrote:
Hey Gang,
I am working on some software for protein structure validation for
structures generated by NMR. I was wondering about using the RECOORD
dataset to use as a reference set for generating Z-scores (I am also using
the "Richardson top 500" for high quality X-ray structures. The RECOORD
web site allows browsing of the re-refined structures and experimental
restraint lists, but I need to download these in bulk if possible. Is
there a way to do this?
Kindest Regards,
g.
--
Gary Van Domselaar, Ph.D.
Genome Canada Bioinformatics
University of Alberta
Edmonton, AB, Canada
Phone: 780-492-5969
*** Details on how to be removed from this list as well as an ***
*** archive of recent postings are available at ***
*** http://mail.bmrb.wisc.edu/bmrblist/ ***
This archive was generated by hypermail 2.1.4 : Thu Mar 02 2006 - 12:57:10 CST