[emboss-dev] Question about sequence input-output - Checked by AntiVir DEMO version
pmr at ebi.ac.uk
Mon Jul 10 17:24:36 UTC 2006
Guy Bottu wrote:
> Dear developers,
> I just installed the patched versions of ajseqread.c and ajseqwrite.c in
> order to prepare for the new release of EMBL. I took the occasion to take
> a look at the supported formats for input and output and was a litlte
> surprised :
Very close to the release (Saturday 15th) ... this is the current status
and unlikely to change.
> - for input the names "embl" and "em" support both the old and the new
> format. For writing "embl", "em" and "emblold" write in old format and
> "emblnew" writes in new format. Is that not a little bit confusing ?
We decided it is more confusing to change "embl" output.
There is a new "emblnew" output for anyone who wants the new format.
If it is popular, we may make it the default for "embl" output in a
> - Worse is the following : for reading "ncbi" and "fasta" read NCBI fastA
> format, "pearson" reads standard fastA format. For writing "ncbi" writes
> NCBI fastA format and "pearson" and "fasta" write standard fastA format.
> I remember that a colleague of mine had troubles because of this. Should
> "fasta" not consistently have the same meaning and then preferably be
> synonymous of "pearson" ?
"pearson" is only there for a few cases where we have to preserve the
"fasta" interprets the ID and can handle the strange NCBI style files.
We need separate output format names because the IDs are so different.
There are a huge number of possible combinations of "fasta" format. we
certainly need to write "ncbi" for some users and "fasta" for others.
> - I also noted that jackkinfernon, nexusnon and meganon are only
> supported for writing, not for reading, even if they are mentioned as
> input formats in
They should not be listed as input formats. If anyone does need them we
can add them in future (so far, nobody has asked for them!)
Hope that helps,
More information about the emboss-dev