EMBOSS: Project Meeting (Mar 29th 1999)
Ian Longden, Richard Bruskiewich
Alan Bleasby, Val Curwen, Thon de Boer, Mark Faller, Sinead O'Leary,
Rodrigo Lopez, Martin Senger
1. Matters Arising
Peter has reviewed the proposed ACD changes and will implement them
after his return from the China ICGEB course. They will make
processing more efficient and simpler, but it is better not to make
the changes and commit them just before going away because ACD
processing is basically fine.
Gary asked about cleaning up the ACD files. This can be done later for
consistency. Most of the things that need to change are workarounds which
are superceded by ACD improvements. The workarounds should still be fine.
Peter will implement the latest additions to the ACD graph processing.
Ian has produced a list of all current graph options.
Alan continues to look into the CVS problems. The "-k" switch proposed
last week made no difference. HGMP will try to set up a test machine
for their client.
Peter will beta test SRS6 and check for compatibility with EMBOSS both
in using getz for database retrieval and in generating Icarus code for
application definitions from ACD files.
2. Library documentation
Library documentation is in two parts, the comments in the source code
which are used for the HTML library manual and the SRS EFUNC and EDATA
databases, and Peter's general introduction with examples.
Peter proposed rearranging the source code files so that these are better
orgnaized and easier to maintain. The header files have "@data" sections
for each data type, listing functions by category. These can be used to
generate "@section" and "@subsection" divisions in the source files,
with the functions sources in order under each subsection.
The same organization can be used in the general introduction.
Some method of including a function in more than one section was suggested.
Peter proposed an "@seealso" section to list them.
Alan suggested adding an index by function name. This can be added to
each HTML page (one per source file) and to a general index page.
Peter requested high priority sections for the introduction. Lists and
Tables were proposed by several people.
3. General progress on release 0.0.4
Alan reported that "make-all-static" is failing on his Linux machine.
Alan has added 5 new applications:
- CpG islands
- CpG islands
- Helical axis representation of proteins
- Helical net representation of proteins
- Coiled-coil prediction for proteins
Gary is working on "compseq" for sequence composition.
Sinead is working on conversion of prosite to a regular expression
4. Requirements for release 1.0
A target date for Release 1.0 was discussed. July was proposed by Peter
as a provisional date, as the ISMB meeting is in August.
Requirements before release 1.0 include:
A general user guide
An Administrator's guide, including database definitions
Some additional sequence formats for input and output
Additional database formats, which have not been thoroughly tested yet.
A few key applications
Peter will make a list on a Web page for review.
5. GFF and Features
Richard asked about applications that could write GFF files for
feature results. He could create a GFF object and GFF output
quickly. GFF input may take longer.
A number of programs identify sequence features, for example "pepcoil"
Other applications read sequence ranges, for example "translate" and "cutseq".
Peter suggested an automatic conversion. If sequence reading could
read and write GFF or EMBL/Swissprot feature tables then seqret would
automatically convert if the input and output formats supported features.
6. Any other business
6.1 WWW interafces
Peter listed the current options for WWW interfaces:
AppLab (Martin Senger, EBI) will use an XML meta data format.
SRS6 will use Icarus application definitions so that applications
can be launched on the results of a query.
www2gcg (Marc Colet, Belgian EMBnet node) uses GCG config files
SeqPup (Don Gilbert) is a Java interface that uses a modified form of GCG
The China EMBnet node has a student working on WWW interfaces. Peter
will visit them next week.
Peter hopes that all of these can be generated automatically from the
AppLab was considered the most important of these.
7. Next meeting
Next Monday is a holiday, so next meeting is on the following Monday
12th April, usual time and place.
Peter Rice, Informatics Division, The Sanger Centre,
Wellcome Trust Genome Campus, Hinxton Hall, Cambridge, CB10 1SA, UK.