Bio::Aln almost ready alpha-version

Georg Fuellen
Wed, 20 Nov 1996 22:27:57 +0000 (GMT)

Hi everyone,
has the newest on the alignment object, Bio::Aln.  PLEASE, DO take a look 
at my questions, reproduced below.

Bio/ now passes a long test program, at
The test program output is at

All other code is at

Bio/ has been improved considerably, too.
It's now got alphabet definitions (are they ok ?),
cleaned up initialization, checking for user errors w/ named parameters,
and various other improvements.
(Can anyone try to fix the dup() function there ? (Steve?))

Please, do send me more test code!

I guess for a public beta-release we'd still need to accomplish most of
the ``to-do'' items for both Bio::Seq and Bio::Aln, and I'm wondering
if anyone could do this for Bio::Seq... (Dag? Lew? Steve? :-)
If you modify code of Bio::Seq, pls tell me about your plan !

best wishes,

#Open Questions
#Q. a: How does one check data input if the data are supposed to be a list
#      of Bio::Seq ? `` if (ref($seq) ne 'Bio::Seq') { ''
#      is certainly _not_ a good idea, because it breaks down
#      if we later on use inheritance (see ``sub _seqs'').
#Q. b: I'd like to support both sequence bags and alignments; the latter have
#      priority. I don't check equal length of the sequences, but shall offer
#      a ``samelength'' flag (currently a function ``samelength'' accesses
#      type[1] which shall be used to hold this flag), and a routine for 
#      checking correctness of this flag. Checking the right-hand-side
#      boundary of a request for an alignment slice then amounts to checking
#      just the length of the first sequence [i.e. ``_width()'' ]; it's assumed
#      that the others have same length, or otherwise "undef" may appear
#      as a letter value. Is this set-up a good idea ? Is there a better
#      alternative ? Can I automagically trigger the appearance of an
#      undef-Symbol (e.g.``_'') in case the array boundary is exceeded,
#      instead of getting ``Use of uninitialized value at'' warnings ?