[Debian-med-packaging] Bug#580277: GMAP -- Align mRNA and EST sequences to a genome

[Shaun Jackman]

Hi Charles,

On 4 May 2010 21:01, Charles Plessy <plessy at debian.org> wrote:
> Le Tue, May 04, 2010 at 06:30:26PM -0700, Shaun Jackman a écrit :
>>
>> I've packaged GMAP:
>> Align mRNA and EST sequences to a genome
>>  GMAP: Genomic Mapping and Alignment Program
>>  GSNAP: Genomic Short-read Nucleotide Alignment Program
>>
>> The ITP is here:
>> http://bugs.debian.org/580277
>>
>> The git repository is here:
>> git://git.debian.org/git/debian-med/gmap.git
>> http://git.debian.org/?p=debian-med/gmap.git
>
> Dear Shaun,
>
> thank you very much for helping us to stay up to date in the field of
> next-generation sequencing. Just out of curiosity, will you use the
> package yourself, or did you prepare it for a friend or a colleague?
>
> I had a look at your package and have the following comments:
>
>  - It is non-free:
>
>   ‘Distribution of this Package as part of a commercial software product requires
>    prior arrangement with the Developers.’
>
>   I updated debian/control accordingly.

Thanks.

> Since this license is not compatible with
>   the GPL, I recommend that you chose another license for your packaging work, in
>   order to avoid headaches in the future.

Can you recommend a license that is DFSG free and compatible with the
upstream license?

>  - We will have to be careful when doing modifications.
>
>   ‘Distribution of a modified version of this Package requires prior
>    arrangement with the Developers.’
>
>   I am not sure of what it means for the manpages you wrote. If you have written them
>   using the output of the programs as a template, they are a derivative of them,
>   and the above may mean that we need agreement of the upstream developpers before
>   redistributing them. But anyway, manpages are better to be forwarded upstream. Have
>   you contacted them?

I'm in contact with the upstream author, but not yet regarding the
Debian package. I'll forward the man pages upstream.

>  - Location of the databases.
>
>   I have seen ‘-DGMAPDB=\"/usr/share/gmap\"’ in the build logs. Does that mean that the
>   users will have to install their databases there, or is there a way to override
>   this path? In general, I think Debian users expect /usr/share to be completely under
>   the control of the packaging system. Perhaps another path would be better?

The default location can be overridden by a command line option (-D)
or an environment variable (GMAPDB).

How about /var/cache/gmap?
http://www.pathname.com/fhs/pub/fhs-2.3.html#VARCACHEAPPLICATIONCACHEDATA

> - Package description. How about this one:
>
> Description: spliced and SNP-tolerant alignment programs for mRNA and short reads
>  This package contains the GMAP and GSMAP programs to align biological
>  sequences, and additional utilities to manage genome databases in GMAP/GSNAP
>  format. GMAP (Genomic Mapping and Alignment Program) is a tool for aligning
>  cDNAs (complementary DNA) to reference genomes. GSNAP (Genomic Short-read
>  Nucleotide Alignment Program) is a tool for aligning single and paired-end
>  sequence reads as short as 14 nt and of arbitrarily long length. It can do de
>  novo discovery of splice sites (including interchromosomal splicing) or use a
>  datbase of known splice sites. It also tolerates known single nucleotide
>  polymorphisms (SNPs) and can align bisulfite-treated DNA.

I edited and reformatted your description. How's this look?

Description: spliced and SNP-tolerant alignment for mRNA and short reads
 This package contains the programs GMAP and GSNAP as well as
 utilities to manage genome databases in GMAP/GSNAP format.
 GMAP (Genomic Mapping and Alignment Program) is a tool for aligning
 EST, mRNA and cDNA sequences.
 GSNAP (Genomic Short-read Nucleotide Alignment Program) is a tool for
 aligning single and paired-end transcriptome reads.
 Both tools can use a database of
 * known splice sites and identify novel splice sites.
 * known single nucleotide polymorphisms (SNPs).
GSNAP can align bisulfite-treated DNA.

>  Maybe it would be good to propose to the authors to proofread this description,
>  so that we are sure that I did not under- or over-estimate the possibilities
>  of their programs.

I believe these claims are correct, but of course it wouldn't hurt to check.

Cheers,
Shaun