What
exactly is BIRCH? - The short answer: BIRCH is a system for organizing large
numbers of bioinformatics programs and databases.
The longer answer: BIRCH is a conceptual framework for building a
bioinformatics
system that is user-oriented, powerful, easy to use, and tailored to
the needs of the local user community. That being said, what you get
when you download BIRCH includes:
A hierarchical
directory structure that organizes programs,
data and documentation
A large set of pre-configured, ready-to-use Open Source
programs for analysis of sequences, markers, genetic data, gene
array
data...
The bioLegato
graphic interface, which greatly simplifies
working
with large datasets, and is particularly good at using output from one
program as input for another program. Think of bioLegato as a program
that
knows how to run other programs.
A database of all documentation
on the system that is readable
through the BIRCH web site
Tutorials that
are task-oriented, showing how to put together
several programs to accomplish a task, rather than focusing on the
details of how to use each specific program. Tutorials include numerous
screen shots with real data. Tutorials cover the critical
theoretical points, possible problems, as well as practial details of
how to accomplish a complex task.
A set of system
administration tools that simplify the process
of making everything work the same way for everyone
Why
should I choose BIRCH, as opposed to EMBOSS or other packages etc...
- You
shouldn't have to choose. Just about any software package can be
integrated into BIRCH. That being said, BIRCH is distinct from other
bioinformatics packages in a number of ways:
The bioLegato interace unifies the access to all programs in a
single graphic interface. Behind the scenes, bioLegato takes care of
interconverting file formats, allowing transparent
pipelining of data from one program to the next
has workarounds for numerous bugs and features that can be
problematic with many of the well-known programs that are in common use
makes it easy to re-do analyses and ask "what if" questions
automatically runs CPU-intensive programs such eg. phylogeny at
a lower priority, preventing these jobs from interfering with the work
of others on the system
adds inormation to output that is not always available from the
programs themselves. Examples would include parameters used during the
run, and execution times for long-running jobs
BIRCH organizes the documentation for many different
bioinformatics packages into a well-organized set of HTML pages.
BIRCH is highly flexible and configurable. It lets you add just
about any program, documentation, or sample datafiles that you want, so
that they appear seamlessly to the user as part of BIRCH and bioLegato.
Can
I install BIRCH on my PC? - Well... yes, if your PC runs Linux.
Otherwise, you have to use BIRCH on a Unix or Linux server. The great
thing about BIRCH, though is that you can access your Unix account (and
therefore, BIRCH) from any computer on the Internet. See Using Unix from Anywhere.
Why
don't BIRCH programs run through the BIRCH web site?
Been there, done that
- There are already more web sites than you can count that let you
paste data into a window and run a program. BIRCH goes way beyond that.
Web pages are ackward
interfaces - For anything but very simple tasks, Web pages are
not a very efficient way to work with data.
the web program runs in your browser. It is not
well-integrated with the rest of your desktop
output is almost always human reable, and not readable by
other programs
you usually have to go to many different web sites to do a
variety of tasks
web pages are usually geared to working with one data item
(eg. sequence) at a time, rather than large datasets.
most web sites have to impose limits on the amount of
processing you can do
web sites are less secure, because your data goes off-site
Why
is BIRCH available for Unix and not for Windows? - The short
answer is that the Windows platform is really specialized for a single
user on a standalone PC. The Windows system is almost impossible to
administer as a secure multiuser system, and multitasking is still
quite unreliable on Windows XP. As well, the Windows interface is
wedded to the "one program owns the screen" method of working. Unix has
always been a multiuser, multitasking system, and Unix desktops have
always been designed to give the user the ability to work
simultaneously in many windows. (Mac users take note: Since OSX is a
full Unix system, plans are in the works for a future BIRCH release for
OSX.)
Is there
a way to install my own programs? - On any Unix or Linux system,
each user can have a directory called 'bin' in their $HOME directory,
in which they can install their own programs. You just need to make
sure that $HOME/bin is in your $PATH. On some systems this is
automatic. Just type 'echo $PATH' to see if a directory called 'bin' in
your home directory is included. If not, you can add a statement like
PATH=$PATH:$HOME/bin
to your .profile file (if you use a BOURNE-type shell like bash) or
setenv
path $PATH:$HOME/bin
to your .cshrc file (if you use a csh-type shell like tcsh). Once your
bin directory is in your $PATH, any program you put into your bin
directory will be found at the command line.
Setting
up your account to use BIRCH
How
do I set up my account to use BIRCH? - You need to run the
newuser script which is found in the directory $BIRCH/admin. For
example, if BIRCH was installed in /home/psgendb, you would type
/home/psgendb/admin/newuser
How
do I turn off my BIRCH access? - Run the nobirch script which is
found in the directory $BIRCH/admin. For
example, if BIRCH was installed in /home/psgendb, you would type
/home/psgendb/admin/nobirch
Does
it matter which shell I use? - A lot of work has gone into
making BIRCH shell-neutral, so you should be able to use any available
shell. BIRCH has been tested using csh, tcsh, ksh, bash and sh. There
is an issue with sh and ksh. It is impossible to get sh to display the
current working directory ($PWD) in the command prompt, and to get that
prompt to update each time you change to a new directory. ksh seems to
be able to do this on Solaris, but I have not been able to get it to
work using ksh on Linux. Therefore, on all systems, I would recommend
against using sh, and on Linux, avoid ksh.
I
ran the newuser script, and now BIRCH works, but none of the global
settings work anymore. For example, I just get a generic prompt like
'>', rather than the systemwide prompt. All shells read a
system-wide rc file that contains commands to set up the shell. After
executing those commands, the shell then looks for an rc file in the
user's $HOME directory, and if it exists, executes those commands as
well. Usually the system-wide rc file is read automatically, but on one
system of which I am aware, the command to read the system-wide rc file
must actually be placed within the user's rc file. Specifically, when
using tcsh, the user's $HOME/.tcshrc file must contain the line
source
/usr/local/common/common.cshrc
It is probably a bad idea, from a system administration standpoint, to
require this. In any case, the only alternative was for the newuser
script not to create a .tcshrc file, which then means that BIRCH
wouldn't work if the user was running tcshrc as their primary shell.
The decision was to document this problem, and, on those rare systems
where one must have such a source line, the user must add it manually.
Sorry about that. I usually manage to come up with a solution that
works everywhere, but I don't think there is a clean solution to this
one. As I said, I have only seen this done on one other system, so I
suspect the problem will be quite rare.
bioLegato
My output
keeps comming up blank - There can be many causes for this, but
the most common is that you forgot to select one or more sequences
before opening a menu to launch a program.
Sometimes
when I try to run a program from the bioLegato pull-down menus, the
program doesn't launch.This
only happens when I'm running a remote
desktop (eg. VNC, SGD). Explanation:This is an artifact of network
latency. When you hold down the mouse over the name of a menu (eg.
Edit)
and hold down the left mouse button, there is a brief time lag for the
mouse-click event to be transmitted to the remote biolegato job, and a
further lag waiting for the result to appear back on your local screen.
Thus, if you try to drag down to the name of a program and release the
mouse, the remote biolegato job doesn't have time to know that you have
released the mouse over a program name. Solution: For remote
sessions, click once on the name of the menu to open the menu. The menu
will stay open. Now, click a second time on the name of the program.
Launchers,
browsers, editors
A program
(eg. mesquite) has asked me to select a web browser or other
application to launch. How do I find it? - You can find the
location of any command at the command line. For exampel, to find the
path to Firefox, type 'which firefox'.
If the output was '/usr/bin/firefox',
then either type in that path, or, if the program gives you a file
chooser, traverse the directory tree to the file.
