[Biopython-dev] Next release plans?

[Peter]

On Fri, Jul 10, 2009 at 1:38 PM, Peter<biopython at maubp.freeserve.co.uk> wrote:
> On Mon, Jun 22, 2009 at 6:57 PM, Peter wrote:
>>
>> Once the beta release is out, we'll resume taking small changes
>> (especially for documentation additions or clarifications) with a
>> view to releasing Biopython 1.51 final in July (probably the second
>> week, after people get back from BOSC/ISMB).
>
> OK, that didn't happen - too much to catch up on at work after
> being away at BOSC/ISMB for a week. Also I will be on holiday
> next week (graduation etc). I will have some limited internet
> access. I'm thinking of doing the final release of Biopython 1.51
> the following week (i.e. the week starting 20th July).
>
> This will be after the annual EMBOSS release, and one little thing
> I want to sort out before we release Biopython 1.51 is mapping
> Solexa/PHRED scores in FASTQ files (specifically what to do with
> a PHRED score of zero which is usually a dummy value, but taken
> literally means "this read is wrong" or "worst than random"). After
> discussion with Peter Rice at BOSC/ISMB 2009, I plan to follow
> his plan for EMBOSS (map PHRED of zero to the lowest used
> Solexa score, -5). Once the EMBOSS release is out, I can use it
> for cross checking our FASTQ conversions.

The FASTQ checking is on going. I have updated our FASTQ
code to map Solexa scores as I understood Peter Rice's
description of the intended EMBOSS behaviour (this is for the
corner case of very poor quality reads). However, due to a
couple of minor bugs I found in EMBOSS 6.1.0 we'll either
have to cross check against their CVS code, or hope they
release EMBOSS 6.1.1 soon.

Cross checking against MAQ would also be worthwhile, but
while there are some patches about to fix a couple of MAQ
FASTQ bugs and include Illumina to Sanger standard
conversion, this isn't in their official repository yet.

I guess I could cross check against BioPerl's new FASTQ
support ...

> Also, we have the Bio.Application.generic_run code to retire,
> which basically means we label it as obsolete and update the
> tutorial to use subprocess (see other thread), but this requires
> cross platform testing.

I still haven't got near my Windows machine to do this. I think
this is important to get done in Biopython 1.51 as we are also
introducing the extended set of command line wrappers.

Nevertheless, a July release is still looking possible. Are there
any other issues that would block the release?

Peter