From Bioinformatics.Org Wiki
[return to Release To Do List]
- revisit how BinaryExists looks for apps. The command string could be any of
open -a appname
- open appname
Custom command doesn't seem to be working on MacOSX
Should View Sequences default to GenBank, rather than FASTA?
blprotein - PXHOM can't work with lowercase amino acids. Where do we do the fix? PXHOM, hom.py, or BioLegato?
Bugzilla #1216 Fix: set readseq to convert aa sequences to uppercase before running PXHOM. Better long term fix is to modify P1HOM and P2HOM to accept lowercase.
When retrieving proteins, output defaults to Textfile. Should default to blprotein.
- Added bltable as a documented BioLegato program. This included
adding to birch launcher
- adding to birchdb
- upgrading of features, including blsort.py
For the special case of building BIRCH on CCL, does it matter whether we run makeframework.csh when BIRCH=/home/psgendb or /home/psgendb/BIRCHDEV? blastdbkit.py sets the BLASTDB variable on a new install, but getbirch needs to make substitutions when installing on a different site. A: makeframework.csh doesn't modify any files, and doesn't reference the $BIRCH environment variable. Therefore, it should make no difference what the value of $BIRCH is.
We should revisit what happens if a file fails to download correctly. Right now, we do not delete the tgz file. If the download fails, should we delete the md5 file? On the plus side, deleting the md5 file could be used as an indication that the file needs to be downloaded, the next time we run an update. That would make the update more of a self-healing process on an incomplete database. However, we need to make sure that the absence of an md5 file doesn't cause other tasks, such as checking the database, to fail.
A: Remove both md5 and tgz files if any part of the download of a database fails. This makes it possible to run an update again, and blastdbkit.py will find files on the remote site that are not present on the local site. The important thing is that the user must inspect the log to know that some files were missed, and to run the update again.
FAQ for BLAST databases
- What if my download stalls? A: There are many reasons why a download may stall. A simple one is that your internet connection temporarily was lost, you ran out of disk space, or the load at the FTP site (eg. NCBI, EBI, DDBJ) was high. One other reason might be that NCBI is in the process of updating the FTP site with new files, which happens with each new GenBank release. If blastdbkit.py tries to download a file as that file is being updated on the FTP site, this will cause blastdbkit to hang. Blastdbkit.py is designed to look for newly updated files each time it runs, so you can always restart the download and blastdbkit.py will resume the download.
- When I run a BLAST search on my local machine, the report pops up in my web browser, but is takes awhile (eg. 10 seconds) before the hits pop up in blnfetch or blpfetch. Why the lag? A: Blast searches run through BioLegato comprise three steps: 1) run the search, and write results in NCBI archive format. 2) Run blast_formatter and generate HTML output to display in the web browser. 3) Run blast_formatter and generate output as a TAB-separated value file (.tsv) to display in blnfetch or blpfetch. Steps 2 and 3 each require several seconds to go back to the database files and collect the information needed for each output file. Consequently, when you see the HTML report pop up in your browser, blast_formatter is still working on getting the report information for blnfetch or blpfetch.
possible multiple concurrent output formats