What happened to my data, or why do I get an Internal Server Error?
Unfortunately, the Biology Workbench is very sensitive to unexpected
conditions, and it often responds by deleting or corrupting the session
in which you were working when the error occurred. This error usually
results in one of two things: an Internal Server Error, or the
session ending up completely empty. In addition, the Biology Workbench
has a physical limit as to the number of user accounts that can be
maintained in an "active" state. If an account has not been
accessed in a while, it will show up as being empty the next time the
user logs in, though the data will still exist in backup.
We can restore all data that existed as of the previous midnight
(approximately). The Restore Sessions tool might be able to recover
your data, if you use it before the next backup occurs. Otherwise, to have
data recovered, please submit a bug report (with the bug report function or
by direct email to
bwbhelp@sdsc.edu), with your
account name and the name of the session you would like restored. If you
are getting an Internal Server Error, it is unlikely you will be able to
fix it yourself, and will need to send a bug report.
To help reduce the chance of these errors occurring or affecting you,
try the following:
- Avoid large sessions: sessions that are larger, that is sessions with
a lot of sequences, or with larger sequences in them. If a session is
too large, the chances of an error occurring increase quite a bit. If
you notice that your session is getting big, then it is time to delete
unneeded sequences, or to create a new session. You can break up a
larger session by repeatedly copying the larger session, then deleting
what you don't need from the copy. We realize this is inconvenient, but
unfortunately it is the only way this can be done.
- Do NOT select sequences after you have used the "Back" button on
your browser: the Biology Workbench stores information in the web pages
themselves, so if you have added or deleted sequences, what is listed
an old copy of that web page (one you have reached by using "Back") is
not going to be correct. The Workbench, in its confusion, may just end
up deleting everything in the session. To avoid this, ALWAYS hit the
"Refresh" button before you select any sequences, or toggle the tool
type (for example, if you are in Protein Tools go to Nucleic Tools,
and then back to Protein Tools).
- Copy your critical data: if you have recently (that is, within the
same day) done a lot of work, copy the session to a backup. Yes, this
does waste a little space, but it adds a lot of safety for you. Just
don't keep around too many copies -- when you make an updated copy
please delete the last "backup" session.
- Don't try to fight a slow network: if the network is very slow,
data errors likely will occur. If you cannot wait, be sure you have
copies of your data.
- Try to keep your Default Session very small (or empty), and use
alternate sessions for doing your work. If any session other than
the Default Session is corrupt, you merely cannot access that session.
If the Default Session is corrupt, you cannot log in to your account
at all.
How safe is my data?
Generally, pretty safe. Our disk is backed up every day or two to tape,
to make restorations from catastrophic disk failures. In addition, we
make local backups of the user session data. Sessions which have been
changed are backed up nightly, and every month all sessions are backed
up. This is why we can restore corrupted sessions in a relatively
short amount of time. Data that has been added since the prior backup
cannot be restored, unfortunately.
However, for many reasons (including portability), we suggest that
Biology Workbench users make local copies of their data. Because our
Session files are not in a portable format, we do not suggest using
"Download Session" for this purpose. Rather, saving the sequences
should be done with the "Download" function or with the "Text format"
link from within the View tool output.
Why isn't TEXSHADE returning a colored alignment?
TEXSHADE crashes quite often, especially for large alignments. Try Boxshade
if you cannot get TEXSHADE to work.
Why isn't the Biology Workbench being improved?
The Biology Workbench currently is not being funded; it is being maintained
by a very small portion of one person's time as a service to the community.
There is a proposal in its initial stages to obtain funding to completely
rewrite the Workbench and use it as a framework to do many more things
than the current Biology Workbench, but even if funding is obtained it
will be a while before it is ready for public consumption.
Why is it difficult or impossible to import large sequences, or
a large number of smaller sequences?
This is because the Biology Workbench was originally designed for work
with protein sequences, and the large (>100K bp) nucleotides create
problems for the Biology Workbench core. An import of a large number
of sequences can also cause the same problems.
Even if the Biology Workbench program could handle large sequences, few of
the tools in the Biology Workbench are useful for genomic analysis, so in
general one should be doing genomic analysis with other services.
How can I draw bootstrap values on the phylogenetic tree drawn
in "CLUSTALTREE"?
Unfortunately, the Drawtree and Drawgram tools from Phylip do not allow for
bootstrap values to be printed on the branches of the trees. One needs to
use their own picture/graphic editor to do this task.
What sort of phylogenetic analysis is done with the Biology
Workbench tools?
The answer is not a lot - implementing all of the Phylip tools would be
a rather big undertaking for us. As of now, we simply use some of the
Phylip tools to display
the neighbor-joining guide trees generated in the Clustal W multiple
sequence alignment program. We offer this functionality as a service to
our users, but suggest that people in need of detailed phylogenetic analysis
look at other services or packages (for example, installing Phylip on their
local machines).
Why isn't the "View Database Records of Imported Sequences (View
Records)" function working correctly?
Occasionally we reindex our databases on different keys, and when that
happens the key that is tied to the sequence in your session is not
correct. Also, sometimes the databases can become corrupt. Try
searching for that sequence and reimporting it if you really need this
feature to work. If it still is failing after you try that, please submit
a bug report.
Why is the Biology Workbench very slow?
This could be due to many factors, but the most likely explanation is that
there is a problem (bottleneck) in the Internet between your site and the
San Diego Supercomputer Center. If you notice this happening quite often,
it probably is worth sending a bug report so that we can see if it is a
problem on our end or on your end.
Why were the sequences I selected not the sequences that were used by
the program I chose?
In most cases, the way around this is to try not using the back button
on your web browser when in the Biology Workbench. Be sure you always
get back to the tool menus by hitting the "Return", "Main", or "Abort"
buttons, and NOT by using the "back" function on your browser.
We store state variables in the web pages, and the location of those state
variables may change. For example, if you import or add
a sequence or alignment, and then use the back button to get back to the
tool menu, the sequences you select on the screen may not be the sequences
that actually are being selected. Using delete in this case can be
disasterous.
There are cases where using the "back" function can be convenient. For
example, if you are running a tool, but simply want to change a few of the
settings from what you had before, you can do this relatively safely (as
long as you do not add, delete, or import during while doing this). We
do have a "Change Settings" button which is a safer way of performing the
same task, though.
Why is the postscript file I downloaded unreadable?
In most cases, we simply provide a postscript file that is generated by
the external application we provide (for example, BoxShade). There is
little we can do to fix those postscript files, though we can contact
the original developers in extreme cases. TeXShade may be a decent
alternative if one cannot get BoxShade to work for them.
Multipage postscript files seem to create problems for many applications.
All we can suggest is trying out various programs, until you find one that
allows you to manipulate the file in the fashion you desire. Ghostscript
(with its viewer Ghostview) generally does a pretty decent job with our
multipage files, and some combination of Adobe tools (e.g. Acrobat) should
probably do the job.
Why does my window hang after I hit "submit"?
There is no simple answer for this, as a multitude of things can cause
this. The most likely cause is that the program you are using is in
progress. Hitting the "Stop" button on your browser and resubmitting
the job usually just floods the Biology Workbench with multiple jobs
doing the same thing.
We have a time-out on the Biology Workbench, that kicks in at a predefined
limit (at this moment, it is 10 minutes). If the job takes that long, the
Biology Workbench kills it, and then resubmits it as a batch job. You do
not need to resubmit it yourself -- it is done automatically. One can use
the "Retrieve Batch Output" function to check on the status of their batch
job.
It is also possible that a slow network connection or an incompatibility
between your browser and the Biology Workbench could lead to a hanging
screen. Please submit a bug report for any persistent behavior along these
lines.
How do "Batch Jobs" work?
In short, one has the option to submit their jobs with "Run as Batch"
selected. When they do that, the screen returns immediately to the tool
menu, and an advisory message is printed. To get the results of a batch
job, one selects "Retrieve Batch Output", and then selects the job they
want (notated by script name and when it was submitted). If the job is
still running, we note that above the list of completed batch jobs.
Note that the batch jobs are separated by what tool type they are
(Protein, Nucleic, Alignment), but they aren't separated by session (I
guess there are pros and cons to this). A user should probably delete
the results of a batch job when they are done viewing it, otherwise the
disk can fill a bit more than necessary.
If somebody has a job take over a certain amount of time (generally
around 5 minutes), they get a message that says that the
job has been killed and resubmitted as a batch job. When that batch job
is complete, they can get the results as described above.
Is my web browser the cause of my problems with the
Biology Workbench?
Often, the answer to this is yes. We are quite dependent on the web
browser used and its settings, and many of the unexplained bugs end
up due to problems on the user end. We discuss recommended web browsers
and settings on our front page, but in essence you probably want to set
your memory cache as high as comfortably possible (over 10 MB if you
can mangage it), and you may want to try turn disk caching off if you
can't solve the problem otherwise.
Who makes your non-redundant protein database?
We do. It is a non-redundant database containing all the protein sequences
that are in the Biology Workbench databases, and gets updated every time our
databases get updated. It is constructed by simple sequence comparison, so
we do not do any elimination of simple sequences or redundant subsequences.
One should exhibit caution when using this database in statistical-dependent
work, but otherwise it works quite well.
Why can't I get MSA to work?
MSA uses a fairly detailed algorithm, which uses enormous amounts of
memory and CPU time for alignments of large numbers of proteins, or for
proteins where the homology is fairly weak (in other words, the proteins
are not very similar). As a result, any MSA jobs that take over 10 minutes
on our system are killed. Clustal W is probably a better choice for doing
the alignment in the aforementioned cases.
Why do I get strange results when searching a Genbank database in
Protein Tools.
Genbank records can contain translations, proteins for which
part or all of the nucleic sequence codes. Many records have no defined
translations at all. In essence, you should probably view the record to
see what can be imported. We label protein-importable translations as
"CDS", and provide a checkbox to define whether or not you want that
translation to be imported.
What is the difference between PDBFINDER and PDBSEQRES databases?
PDBSEQRES has been created based on the PDBFINDER database records, in which
the sequences are derived from PDB. The database contains only those records for
which the ID is found both in PDBFINDER database and in PDB SEQRES file. PDB
SEQRES contains the updated listing of all PDB sequences in FASTA format.
PDBFINDER records return sequences that appear to be derived from the atomic
coordinates (and therefore often fragmentary) and do not provide information on the
gaps due to the unresolved residues. We replaced those sequences with the corresponding,
complete ones from the PDB SEQRES text file. PDBSEQRES therefore has complete, updated
sequences from PDB. Everything else is left as in the original PDBFINDER record. The
complete sequences are especially more useful in the case of viewing protein structures.
What happened to the "Close Session" button?
In version 3.2 of the Biology Workbench, the user is always in a session.
If they don't specify a session, they are in a special session called
"Default Session" (these changes make the Biology Workbench
more efficient, and reduce user data loss). Closing a session is the same
thing as resuming the "Default Session", so it was a redundant function.
