[Bioperl-l] latest bioperl commit
Nathan Liles
nml5566 at gmail.com
Mon Feb 22 20:33:43 EST 2010
I just made a major commit to the bioperl svn. These changes all affect
the genbank2gff3.pl converter. The following is a brief synopsis of the
changes:
New command line options added:
/--noinfer /: don't infer exon/mRNA subfeatures
For prokaryotes, sometimes having the converter infer these
subfeatures from genes is less than desirable. This will turn that off.
/
--sofile=so.obo /: obo file used to create up-to-date type mappings
If you pass a so.obo file to the converter, it will try to
match every primary tag with a valid SO ID. So, for nebulous tags like
misc_feature, it will try give it a more descriptive name based on the
note field. It will also validate parent/child SO relationships and ID
uniqueness. If you don't have a local file, you can use --sofile=live to
grab the latest file from sequenceontology.org.
/--manual /: manual curation
Since the converter will likely do a terrible job guessing
tags, passing this flag will allow you manually choose the proper term
for a GenBank entry. It will give you a list of best matches, search SO
from the command line, and allow user input when all else fails. This
flag only works in conjunction with --sofile
/--conf/ : path to config file
Because no one wants to manually annotate a GenBank file every
time they convert it, and because even the best matches can't beat a
human curator, the converter also has the option to save manual curation
preferences to a conf file. Any preferences saved to a conf file will be
automatically applied to the GFF3 everytime you convert it. The conf
file consists of "rules" for when to match a primary tag to a GenBank
entry. The generated conf file will be in YAML format. So, if you are
comfortable with the format, it should be easy to manually create/edit
the file. If you use the --manual flag and input a term, it will ask you
if you want to save your preference to the conf file. At the moment,
there is no conf file documentation.
I'm always open to hearing any feedback/suggestions/bug reports.
-Nathan Liles
More information about the Bioperl-l
mailing list