[Gusdev-gusdev] conference call summary

[Chris S. <sto...@SN...>]

Conference call summary July 23, 2002
CBIL: Jonathan Crabtree, Sharon Diskin, Yongchang Gan, Li Li, Joan 
Mazzarelli, Debbie Pinney, Angel Pizarro, Jonathan Schug, Chris Stoeckert

PSU: Matt Berriman, Christiane Herz-Fowler, Keith James, Paul Mooney, 
Christopher Peacock,
Marie-Adele Rajandream, Adrian Tivey

Agenda:
	*	time scale and priorities for completing GUS 3.0
Jonathan C. reported that we will be migrating GUSdev to GUS3.0 over the 
next 3-4 weeks. Once data is migrated, then testing of plugins can be 
done. The web interfaces need to be adjusted to use GUS3.0 too. AllGenes 
has been made GUS3.0 compliant but obviously has not been tested. At 
issue is the ability to query GUS through GeneDB and to be able to 
demonstrate this at the upcoming Woods Hole meeting in September. 
Because of the uncertainties involved in bringing the new system online, 
it may be best to implement the current GUSdev for which all the plugins 
and interfaces are available. We need to make the current schema and 
code available for GUSdev. Note that RAD3 (in GUS3.0) is currently being 
populated along with supporting tables in CORE and SRES. Also note that 
the next public releases of AllGenes and PlamoDB will be with GUSdev and 
that public views of GUS3.0 probably won't occur until November/December.
ACTION ITEMS: JC- generate CREATE TABLE statements of current GUSdev. 
CBIL: Update files as necessary at 
http://www.cbil.upenn.edu/downloads/GUS/

	*	our suggested new tables protein properties and protein 
features, did you
guys have any comments on this for us ?
The ProteinProperty table and protein features views of AAFeatureImp 
proposed by Arnaud by email are OK with CBIL. The views can be used with 
the Interaction table to specify interactions of sequences. Note that 
interactions between proteins (at the central dogma level) can also be 
made.

	*	availability of up-to-date Plugins for loading of the controlled
vocabulary tables eg for GO, Taxon into GUS 3.0
Jonathan S. reported that it should only take a day or two to adjust the 
current plugins to work on loading GUS3.0. No immediate plans to do this 
but if needed can schedule it.

	*	availability of Plugin for populating EMBL and Swall into External
NASequence
Angel reported that current plugins for populating sequence and feature 
tables from GenBank flat files can be adapted to use EMBL and SWall 
files. The current plugin just needs to call parsers for these (which 
need to be written).

	*	Additional keys and qualifiers for Artemis to simplify 
transfer of data
into GUS / GeneDB
It was noted and agreed that all keys and qualifiers should make use of 
controlled vocabulary/ontology tables in SRES (e.g. SequenceOntology).

	*	Annotation interface
Joan reported that the CBIL group had gone over Artemis and Apollo but 
will probably create their own stand-alone annotation tool using 
whatever parts of code they can borrow from these. At issue for CBIL was 
the need for various accessory programs (from BLAST to Perl tk) for 
Artemis to be installed locally. CBIL's annotators are distributed 
around the world and need to use accessory programs (like BLAST) over 
the web. PSU and CBIL have overlapping requirements and will continue to 
share ideas and hopefully code.
ACTION ITEM: SD send latest requirements document out. (this was done 
but stuck due limit in sourceforge mail size).

	*	Biojava and GUS Central Dogma objects
Paul reported that work is proceeding on the central dogma objects but 
they are not ready to release yet.
ACTION ITEM: PM send BioJava for GUS objects to JC as is to give him a 
preview.

	*	Phenotype tables
SeqVariation view of NAFeatureImp can be linked to Phenotype table 
through a linking table SeqVariationPhenotype. This allows changes at 
the genomic level to be linked to phenotype.
ACTION ITEM: CS create Phenotype and associated tables in SRES.

	*	Population ortholog tables
Li described population of ortholog tables. There are two parts: a 
general part for loading sequences, BLAST comparisons, and loading the 
Similarity table; and a specific part for using the similarities to 
create ortholog and paralog groups. Li is using reciprocal best hits 
combined with a graph clustering approach. Plugins for the general part 
are available. Ortholog tables for proteins are present in GUSdev; these 
have been expanded in GUS3.0. Keith raised the issue of curating 
orthologs. PSU will look into populating ortholog tables by curators 
directly.
ACTION ITEM: LL will make her code for populating ortholog tables 
available by September.

	*	CVS
Sanger CVS relies on Sanger administrators but are very likely to be 
more responsive than Sourceforge administrators. Sanger does not have 
trackers but this is not something of interest. Sourceforge does have a 
nice web site for browsing CVS and mail archives. Sanger also has a web 
view for read-only. Will continue with Sourceforge for now but will 
continue to look into Sanger site.
Step releases were thought to be useful by both teams.  Paul suggested a 
naming convention for tagging the CVS tree or the release. A step 
release would cover schema/perl/java/html etc. so one could for example 
check out PlasmoDB as of Release-123-01-Apr-2002.  It is not an 
immediate requirement but something to consider for the future.
ACTION ITEM: PM send URL for Sanger web view of CVS. Find out if 
available for read write.