Bugzilla Bug 179327

I would appreciate if all supporting arches (x86, ppc) would stabilise
sci-biology/t-coffee. It has been in Portage for a while and works well.
Sources for the latest current stable version are no longer available, so I
want to remove that version ASAP.

Test instructions:

After compiling t-coffee, copy the "sample_seq1.fasta" file from
"/usr/share/t-coffee/example" to your home directory. This file contains the
amino acid sequences for four homologs of a protein in different species.

Run "t_coffee sample_seq1.fasta" to have t-coffee align the sequences,
calculate the alignment scores, construct a dendrogram, etc. Output files are
created in the working directory.

You can check the validity of the calculations by comparing the file
"sample_seq1.aln" with the file I am attaching. (This output file is also
available in the example directory, but it was created with an old version of
the program, and the formatting is a bit different so you would have to compare
the files manually. With the provided file, a diff should suffice.) The only
difference between the files should be the CPU time used for the calculations.

Thanks in advance,

Created an attachment (id=119904) [details]
Reference sequence alignment

x86 stable

I'm seeing some differences between the output of t-coffee and your file:

diff sample_seq1.aln sample_seq1.aln.bugzie               
1c1
< CLUSTAL FORMAT for T-COFFEE Version_4.96 [http://www.tcoffee.org], CPU=0.17
sec, SCORE=34, Nseq=4, Len=81 
---
> CLUSTAL FORMAT for T-COFFEE Version_5.05 [http://www.tcoffee.org], CPU=0.08 sec, SCORE=33, Nseq=4, Len=81 
2a3,4
> hmgl_trybr      kkdsnapkramtsfmffssdfrskhsdls-----ivemskaagaawkelgpeerkvyeem
> hmgt_mouse      -----kpkrprsayniyvsesfqeakdds-----aqgklklvneawknlspeekqayiql
4d5
< hmgl_trybr      kkdsnapkramtsfmffssdfrskhs-----dlsivemskaagaawkelgpeerkvyeem
6,7c7
< hmgt_mouse      -----kpkrprsayniyvsesfqeakddsaqgkl-----klvneawknlspeekqayiql
<                       ***. ::: .: ..  .. .             *  .  *: *    :  :   
---
>                       ***. ::: .: ..  .. .  .          *  .  *: *    :  :   
9d8
< hmgb_chite      aatakqnyiralqeyern-gg
11d9
< hmgl_wheat      anklkgeynkaiaaynkgesa
12a11,12
> hmgb_chite      aatakqnyiralqeyern-gg
> hmgl_wheat      anklkgeynkaiaaynkgesa

beer >> coffee; sending back to maintainer... :P

confirmed on an another G4. ribosome?

This should not happen, so I think it is best for version 4.96 not to be
stabilised on PPC. These variations could be caused by the magnification of a
small number-rounding difference. The differences are very small, though, and
probably do not impact the alignment's validity, so I do not think it is
necessary to mask t-coffee in the PPC profiles.

This leaves the problem of missing sources for the last (and only) stable PPC
version. I suggest dropping the stable keyword and having t-coffee available
only in the testing branch (for PPC).

What do you PPC folk think?

(In reply to comment #6)
> This leaves the problem of missing sources for the last (and only) stable PPC
> version. I suggest dropping the stable keyword and having t-coffee available
> only in the testing branch (for PPC).
> 
> What do you PPC folk think?

That's okay for us.  Just drop the stable-version and keep the current version
~ppc.

So, let's close this bug.