[Debian-med-packaging] Bug#601568: bwa: Bug in scoring matrix incurs errors in indel calling.
Charles Plessy
plessy at debian.org
Wed Oct 27 12:02:53 UTC 2010
Package: bwa
Version: 0.5.8a-1
Severity: important
As discussed in the upstream mailing list, bwa 0.5.8a has a scoring
matrix that produces some scientifically wrong results. This impairs
some uses of the program, but not all.
http://sourceforge.net/mailarchive/forum.php?thread_name=alpine.DEB.1.10.1006151606060.21551%40deskpro19670.dynamic.sanger.ac.uk&forum_name=bio-bwa-help
The scoring matrix has been corrected in version 0.5.8c, that I will
upload shortly. This version contains a couple of other changes:
ChangeLog | 43 ++++
INSTALL | 9 +
Makefile.am | 13 +
autogen.sh | 2
bwa.1 | 2
bwa.txt | 452 ----------------------------------------------------
bwape.c | 2
bwt_gen/Makefile.am | 4
cleanup.sh | 1
config.h.in | 157 ++++++++++++++++++
configure.in | 49 +++++
debian/changelog | 12 +
debian/control | 2
debian/copyright | 115 +++++++------
m4/acx_pthread.m4 | 271 +++++++++++++++++++++++++++++++
main.c | 2
sracsfq2fq.pl | 14 +
stdaln.c | 10 -
18 files changed, 649 insertions(+), 511 deletions(-)
(debdiff attached).
As a scientist, I do not feel like distributing bwa 0.5.8a in
Squeeze. Likewise, I would not feel like backporting the correction of
the scoring matrix to 0.5.8a, because the changelog metions other bug
fixes, and I do not want to take the responsibility to distribute a
modified version from which these corrections would be excluded.
If it is not possible to replace 0.5.8a by 0.5.8c (which is completely
understandable and accepted in advance), then I would recommend to
simply remove it, especially that we do not distribute bwa in Lenny.
Have a nice day,
--
Charles Plessy
Debian Med packaging team
http://www.debian.org/devel/debian-med
Tsurumi, Kanagawa, Japan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bwa_0.5.8a-1_0.5.8c-1.txt
Type: text/x-diff
Size: 56919 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/debian-med-packaging/attachments/20101027/3733e6c6/attachment-0001.diff>
More information about the Debian-med-packaging
mailing list