[Bioperl-l] Re: [Bioperl-guts-l] Notification: incoming/888

Hilmar Lapp hlapp@gmx.net
Mon, 29 Jan 2001 23:10:26 -0800

bioperl-bugs@bioperl.org wrote:
> Generic Features created from a GFF string do not
> record Frame information, and when dumping the feature
> out as GFF it is invariably reported as frame = 0.
> The problem is multi-fold:
> (1) the _from_gff_string and _from_gff2_string
> subroutines in Generic.pm do not contain any code to handle the
> recording of Frame information in the feature object
> (2) GFF allows a "." as the frame (meaning info not available),
> while $Feature only allows 0,1, or 2.  Thus it isn't clear how a
> GFF frame of "." should be recorded.  My first thought was that a
> value of undef might return "." in a call to SeqFeatureI::gff_string,
> however...
> (3)  ...it appears that even if there is no frame information
> available in a Feature object, it nevertheless passes the
> $Feature->can('frame') test in SeqFeatureI::gff_string
> and returns a (default??) value of 0 for the $Feature->frame call
> (though there *is* code there to assign the frame to
> "." if it fails the ->can test...)
> I am willing to fix this problem myself, but I would appreciate having
> a consensus from the group about which level of the problem needs to be
> fixed to keep everyone else's code  happy.

I think that frame information should be consistent between GFF
representation and object representation. '.' is equivalent to
undef, and otherwise the frame should be 0, 1, or 2, regardless of
object or GFF string.


Hilmar Lapp                                email: hlapp@gmx.net
GNF, San Diego, Ca. 92122                  phone: +1 858 812 1757