Bio::FeatureIO gff3 fails to parse canonical GFF3 file from GFF3 spec #8
Description
Author Name: Keith James (Keith James)
Original Redmine Issue: 2429, https://redmine.open-bio.org/issues/2429
Original Date: 2008-01-03
Original Assignee: Bioperl Guts
I was getting this error from all my GFF3 files:
——————- EXCEPTION: Bio::Root::Exception ——————-
MSG: Oops! ID cds00001 exists more than once in your file!
STACK: Error::throw
STACK: Bio::Root::Root::throw /software/perl-5.8.8/lib/site_perl/5.8.8/Bio/Root/Root.pm:359
STACK: Bio::FeatureIO::gff::next_feature_group /software/perl-5.8.8/lib/site_perl/5.8.8/Bio/FeatureIO/gff.pm:206
Several code comments within Bio/FeatureIO/gff.pm indicate that it assumes an unique ID for each line. The spec states that an ID must be unique within a file, but one assumes that this means unique /per feature/ where a feature may be split over several lines. The example GFF in the spec supports this interpretation.
To confirm the bug I created a copy of the canonical GFF3 example from the spec and reproduced this error (see above); the cds00001 feature is split over 4 lines, each with the same ID.
This is with bioperl release bioperl-1.5.2_102 on Linux i686, Perl v5.8.8