MAGE Jamboree 2003

1-6, December 2003
Wellcome Trust Genome Campus, Hinxton, Cambridge, UK

Introduction

The goal of this jamboree is to develop 'best practices' of how to encode MIAME in MAGE in a consistent manner to ensure full interoperability of MAGE-ML based tools, i.e. to make sure that the files generated by different MAGE-ML supporting tools can be interpreted by all other tools correctly. In particular we would like to establish fully automated pipelines from various MAGE-ML writing or reading tools to ArrayExpress public archive for microarray data. As a part of this activity, we may wish to define the minimal set of MAGE resources sufficient for encoding MIAME information.

Although this will be primarily a "file exchange jamboree", not a "programming jamboree", we would like to achieve the uniformity of MAGE-ML files during the jamboree for the currently functioning MAGE-ML tools (as much as it is possible because of the limitations of accessing external databases remotely from the EBI). We will concentrate on solving the problems where the collective mind is the most important: design, setting community-wide principles and new ideas for future.

We should care about uniform data encoding mostly where queries/data mining is important. Uniformity may not be so important if MAGE structures are used to document how somebody performed an experiment just for documentation sake, to be read mostly by humans.

The 'jamboree' will take place in the Wellcome Trust Conference Centre (Hinxotn, near Cambridge UK) training room, which is equipped with about 25 Internet connected PCs. The concrete agenda of the meeting will depend on the exact list of participants. There may be three major tracks:

1. for those who are already have working MAGE-ML exporting or importing tools (e.g. Stanford, SMD, Affymetrix),
2. those who have early versions of MAGE-ML functionality, and, possibly,
3. those who only are at the beginning of adding MAGE-ML functionality to their tools. In case the track 3) will be necessary, we will start with a MAGE tutorial.

In addition to the invited participants from organisations that have already demonstrated MAGE-ML functionality for their tools, the jamboree attendance will be based on first come - first serve basis. There will be a cost-recovery registration fee for the participants, covering the costs of the training room and lunches. Additionally the accommodation can be booked in the campus facilities.

Draft Agenda

0. MAGE tutorial - if there is need for it (depending on the participation).
1. Find out what the ambiguous places in MAGE are and how people model their data:
   - Those who have working MAGE export/ concrete plans for it - what decisions have been taken - short presentations (it is encouraged to make examples available prior to the meeting).
   - ArrayExpress and others who are interested in MAGE-ML import - what do they need - short presentations
   - Policies for using the MGED Ontology as a Database/DatabaseEntry and for OntologyEntry will be presented
   - Discussions on MAGE-area-by-area basis
   
   Some of MAGE areas worth considering:
   - ExperimentalFactor usage
   - BioMaterial - Treatment cycle
   • what are important properties of BioMaterials
   - BioAssay usage
   • structure
   • links to Experiment
   • links to ExperimentalFactors
   - ArrayDesign
   • how Features/Reporters/CompositeSequences are used
   • their grouping
   • DatabaseEntry/OntologyEntry usage
   • what are important properties for data mining
   - QuantitationType usage (Special/Standard, ...)
   - OntologyEntry
   • adopting the MGED Ontology policies
   - The use of ADF files (although formally this is not a part of MAGE, we know that many use them and we should discuss if and how to incorporate them, or some other lightweight array design description framework, in the future MAGEv2)
2. Work on MAGE-ML export/import (concentrating on design, coding if there is time).
3. For those who are happy with their export/import - further work on MAGEstk.
4. Sum up what has been done
   - data producers - changes made/planned
   - data consumers - modified assumptions on how MAGE looks like
   - "minimal" MIAME-to-MAGE mapping
   - ideas for MAGEv2, SystemsBiology-OM, Proteomics-OM