[emboss-dev] EMBOSS seqret FASTQ support

Peter biopython at maubp.freeserve.co.uk
Fri Jul 24 09:59:52 UTC 2009


On Thu, Jul 23, 2009 at 5:24 PM, Peter Rice<pmr at ebi.ac.uk> wrote:
>
> Peter C. wrote:
>>
>> I've got another issue for you, which I think is an rounding problem
>> converting negative Solexa scores into ASCII (which sounds a bit
>> strange), or assuming you store everything as PHRED scores in
>> memory, this could be in how you round negative Solexa scores
>> on conversion back to ASCII.
>
> Yup, it's the rounding on output. It was adding 0.5 and going to the
> nearest integer.
>
> For negative values of course it has to subtract 0.5 to get the correct
> rounding.

C can be fun like that - nearest integer verses truncation to lowest integer.

I'd like to re-test with your fixes. I presume these things are being fixed
in the public CVS repository, so I could try building EMBOSS from there.
Is there a particular branch? Or are you planning an EMBOSS 6.1.1
release shortly?

Thanks,

Peter



More information about the emboss-dev mailing list