Procedures for retrieval and installation of AAtDB
Mike Cherry 726-5955
CHERRY at Frodo.MGH.Harvard.EDU
Fri Jul 10 18:04:16 EST 1992
This message describes how to obtain and install AAtDB, the Arabidopsis
genome database funded by the USDA Plant Genome Research Program through
the National Agriculture Library and maintained by a group at Massachusetts
General Hospital and Harvard Medical School.
We have prepared a user manual/tutorial for AAtDB that includes examples
of basic procedures and pictures illustrating how to use the ACeDB
software. We feel it is ideal for scientists who have never used a Unix
workstation. Please reply to this message with your postal mailing
address and we will send the manual free of charge. Please note that we
will give first priority to Arabidopsis groups that are using the
You can retrieve the AAtDB database and the ACeDB software from
weeds.mgh.harvard.edu (IP 18.104.22.168) using anonymous ftp. The
username is "anonymous" and the password is simply your mail address.
If you are not familiar with ftp the procedure is given below.
To use AAtDB on a Sun Microsystems SPARCstation (any model using
OpenWindows) you will need to retrieve only two of the files mentioned
below, that is the bin.at-sparc_1-1.tar.Z and INSTALL.
To use AAtDB on a DECstation (Ultrix), NeXt, or Silicon Graphics
workstation you will need to retrieve the ACeDB executable file from
ncbi.nlm.nih.gov (IP 22.214.171.124) via anonymous ftp, more on this below.
Plus you will need the update.at_1-1.tar.Z and INSTALL file from
For all systems you should retrieve the INSTALL csh script from
weeds.mgh.harvard.edu, NOT the one from ncbi.nlm.nih.gov. This script
will perform the installation for you. We suggest that you create an
account for the manager of the database as well as an account for the
users of the database. This way users will not accidently damage the
database. Into the home directory of the database manager account, for
example /files/aatdbmgr, place all the compressed tar files (those files
mentioned above ending in "tar.Z"; tar is a Unix archiving utility and
compress is a utility that reduces the size of files with binary
encoding) and the INSTALL script. You should not uncompress these tar
archives, let the INSTALL script do that. If you have uncompressed the
tar archives the INSTALL script will compress them again so that the
AAtDB files can fit under 50 Mb.
Now in the home directory of the AAtDB manager (/files/aatdbmgr
mentioned in the above example), with the required compressed tar
archives and the INSTALL script in place, issue the following two
% script aatdb.log
% source INSTALL
The first command will create the file aatdb.log of all the messages
passing over your screen so that you can later review the installation
procedure. The second command installs AAtDB into the current
NOTE: THE INSTALL SCRIPT WILL EXTRACT THE FILES IN THE COMPRESSED TAR
INTO THE CURRENT DIRECTORY. THUS YOU SHOULD BE IN THE DIRECTORY THAT IS
TO CONTAIN THE AATDB FILES. WE STRONGLY SUGGEST THIS BE A SEPARATE
DIRECTORY FROM ANY OTHER USED.
After installation you should minimally have the following files and
directories with these permissions:
The INSTALL script will create the aatdb and textaat csh scripts. You
can then relocate these files to a common directory on your system, for
With the release of ACeDB software version 1.7 we are now using the
executable as it comes from Richard Durbin and Jean Thierry-Mieg. Thus
if you are running ACeDB and AAtDB you only need to keep one copy of the
executable files in the bin directory. However all the files in the
database and wspec directories should be considered specific to AAtDB.
The ACeDB 1.7 software also includes a text-only version of the
database that uses the query language. All information in the
database is available via the text version. A script for textaat is
created during the installation. The query facility is very powerful
but the interface is in its infancy. At our site we have modified this
text version for use as a network server. To connect with the network
service you just need access to telnet on any internet connected
computer. For Unix systems the command on your local computer is:
telnet weeds.mgh.harvard.edu 7001
On VMS systems it is probably something like:
If you want the source to the ACeDB software it is available from
ncbi.nlm.nih.gov via anonymous ftp.
There are three files available on weeds.mgh.harvard.edu via the
anonymous ftp account.
bin.at-sparc_1-1.tar.Z This file contains the executable, binary
database, wspec (models) files, and
documentation files. This and the INSTALL
file are the only files needed if you will
be running the database on a SPARCstation.
update.at_1-1.tar.Z This file contains the text update file,
wspec (models) files and documentation
files. This file and the INSTALL file are
needed by sites which are NOT using a
SPARCstation. In addition, you will need
to obtain the executable file from
ncbi.nlm.nih.gov as described below.
INSTALL csh script that extracts all files from
the compressed tar archives, sets file
permissions and creates scripts to start
aatdb and textaat.
If you need the executable for a DEC, SGI or NeXT workstation, or you
want the source to the ACeDB software you can retrieve these via
anonymous ftp from ncbi.nlm.nih.gov. Look in the repository/acedb
directory for the following files:
bin.mips.1_7.tar.Z ACeDB bin, wdoc, and wspec directories
for DECstation 3100, 5100, etc.
bin.next.1_7.tar.Z ACeDB bin, wdoc, and wspec directories
for NeXT stations.
bin.sgi.1_7.tar.Z ACeDB bin, wdoc, and wspec directories
for Silicon Graphics IRIS workstations.
source.1_7.tar.Z ACeDB software version 1.7 C source requires
GCC and MIT X11R4 libraries to compile.
A note to those obtaining the bin tar archives from NCBI. These tar
files contain more than just the ACeDB software. The wspec directory
files are those specific for ACeDB, i.e., for the C. elegans database.
AAtDB requires different wspec files, so if for some reason you
install these files by hand and not by using the INSTALL script be
sure to extract the files obtained from weeds.mgh.harvard.edu last.
That way the AAtDB wspec files will overwrite the ACeDB wspec files.
The first time you run aatdb on a DEC, SGI or NeXT workstation you will
be prompted to create a new database file. Answer yes or continue to
these questions. You will then need to select "Write Access" from the
main windows pop-up window by pressing and holding the rightmost mouse
button. Using the same procedure selection "Add Update". Then again
from the main windows pop-up menu select "Save". Now in the "Update"
window click on the "Update" button with the leftmost button. The
update process will automatically read the update file and perform
several tasks that builds the binary database file. When asked to
include the allocation for the database file initially try 25000.
An example ftp session to weeds.mgh.harvard.edu:
% cd /files/aatdbmgr move into the directory
that will be the home
for the AAtDB database
% ftp weeds.mgh.harvard.edu Start ftp to weeds
Name (weeds.mgh.harvard.edu:jmc): anonymous enter the proper username
Password: enter your email address
ftp> cd aatdb move into the aatdb
ftp> get INSTALL retrieve the INSTALL
csh script in ascii mode
ftp> binary binary, turns on binary
mode of transfer
ftp> get bin.at-sparc_1-1.tar.Z retrieve compressed tar
archive for SPARCstation
ftp> quit Quit ftp session
I believe that should be enough information for you to set up AAtDB in
under an hour on your workstation. Please review the ACeDB
documentation that will be placed in the wdoc directory during the
If you do not have access to the Internet for FTP retrieval of the
database files you might look into using ftpmail at decwrl.dec.com. This
is a service that that will retrieve a file via ftp and then mail it
back to you. Note that the compressed tar archive files are very large
binary files. Ftpmail must be told that the file is binary. Plus you
must specific the type of encoding to use, binary files cannot be sent
via electronic mail unless they are encoded in an ascii form. UUENCODE
is a standard Unix encoding program. This procedure is not for the
faint at heart as ftpmail will split the uuencoded file into small
pieces. The SPARC binary archive will be retrieved in about 215 50K
pieces of the uuencoded tar archive. Each of these 215 messages would
then need to be edited. Finally the 215 messages are concatenated
together and run through uudecode to produce the 7 Mbyte compressed tar
archive. For more information on ftpmail send a message with just the
word "help" (without the quotes) as the body of the message to the
address ftpmail at decwrl.dec.com.
Please feel free to contact Sam or Mike via email or FAX with your
questions and suggestions.
A general address that is monitored by all the AAtDB staff is:
curator at frodo.mgh.harvard.edu
Mike Cherry Sam Cartinhour
cherry at frodo.mgh.harvard.edu cartinhour at frodo.mgh.harvard.edu
FAX: 617-726-6893 FAX: 617-726-6893
More information about the Bionews