[Bioperl-l] tcoffee tests in BioPerl

[Heikki Lehvaslaiho]

Peter and Jason,

cc: Cedric

To really start using Bio::SimpleAlign, I installed clustalw (v. 1.81)
and TCoffee (v. 1.34).  No problems there; tests provided with the
programs pass. All the clustalw bioperl tests passed (after I fixed
the nse names), too.

TCoffee bioperl tests, however, complain even after I've changed the
expected nse to "id/start-end":

1..13
ok 1
ok 2
ok 3
ok 4
ok 5
ok 6
not ok 7
# Test 7 got: 'CATH_HUMAN/1-335' (t/TCoffee.t at line 86)
#   Expected: 'CATL_HUMAN/1-333' 
    ( failed tcoffee profile alignment using input file
CATH_HUMAN/1-335)
not ok 8
# Test 8 got: 'CATH_HUMAN/1-335' (t/TCoffee.t at line 97)
#   Expected: 'CATL_HUMAN/1-333' 
    (failed tcoffee profile alignment using SimpleAlign input
CATH_HUMAN/1-335)
ok 9
ok 10
ok 11
ok 12
not ok 13
# Test 13 got: 'CATH_HUMAN/1-335' (t/TCoffee.t at line 113)
#    Expected: 'CATH_RAT/1-333' 
     (failed adding new sequence to alignment CATH_HUMAN/1-335)



Tests 9..12 are new and show that the new sequence really get inserted
into the alignment. By printing out data structures, I've verified
that profile alingments work, too.

It seems to me  that since version 1.34 is most probably newer than
versions used when the  test were created, the way TCoffee orders
sequences in the alignment might have changed. In fact, I did not find
from the program documentation anything that would make me expect that
the output sequences should be in any particular order. I'll cc: this
mail to Cedric to get his comment on this.


If only the order has changed, the percentage_identity() of the
alignment should be the same even in older versions. Tests 10 and 12
give the integer values 41 and 47 before and after adding the new
sequence.

Jason/Peter, could you do 'cvs update' and see if these %identity
values hold even with the t_coffee versions you are using?

I ran the clustalw and tcoffee tests under bioperl-07 which I have not
updated and got identical results. It was a relief to find get an
extra confirmation that I have not messed up anything in the bioperl
alignment code.

If everything is as I suspect, it is easy enough change the tests from
checking a name of the sequence in a given position in the alignment
(which conflict the data encapsulation principle!) to comparing the
number sequences in the alignment.


        -Heikki


-- 
______ _/      _/_____________________________________________________
      _/      _/                      http://www.ebi.ac.uk/mutations/
     _/  _/  _/  Heikki Lehvaslaiho          heikki@ebi.ac.uk
    _/_/_/_/_/  EMBL Outstation, European Bioinformatics Institute
   _/  _/  _/  Wellcome Trust Genome Campus, Hinxton
  _/  _/  _/  Cambs. CB10 1SD, United Kingdom
     _/      Phone: +44 (0)1223 494 644   FAX: +44 (0)1223 494 468
___ _/_/_/_/_/________________________________________________________