EMBOSS: Project Meeting (Apr 10th 2000)


Sanger Centre: Peter Rice, Ian Longden
HGMP: Gary Williams, Tim Carver
Apologies: Alan Bleasby, Val Curwen, Jon Ison

1. Matters Arising

2. General progress on release 0.0.4

Peter has fixed the bug in "revseq" which would affect all reverse complement calls with ambiguous base codes.

Peter has replaced "seqret" with "seqretall" but added a new "firstonly" option to return a single sequence for sites who depend on only getting one sequence from a run. "seqret" will be added as a "make check" application.

There was discussion about feature coordinates. Should the start and end positions refer to the original sequence or to a selected subsequence? No firm conclusion was reached, especially for output feature tables.

Peter will propose a new validation procedure for sequence ranges with a link to a sequence object and a default range of "all" to replace the present method which is causing complications when -sbegin and -send are specified together with a range.

Peter will propose new attributes and qualifiers to control whether output files are created and/or old files deleted.

Gary would like an application to align two very similar long sequences (about 50kb each). supermatcher is not enough, as it has a limited number of gaps allowed. Gary is looking at sim3 and suffix tree solutions. The output should be a feature table of differences between the two sequences.

3. Graphics

Peter is still working on the colour map problem in X11 graphics. The cause is the colour map being completely filled (e.g. by StarOffice) and PLPLOT failing to reallocate colours.

4. Documentation

In progress.

5. Any Other Business

Peter will visit Lion UK in Cambridge next week to discuss integration of EMBOSS under SRS 6.0.6

6. Next meeting

Meeting on Monday 17th April, 11:00am, usual place.