The Microarray Gene Expression Data society, which was responsible for developing the microarray community's MAGE-OM object model and MAGE-ML data format, has set its sights on a broader target for its next round of work.
Recognizing that MAGE-OM has many "generic" features that could be extended beyond microarray data — and with an eye toward the rise of cross-platform experimental protocols — the next version of MAGE-OM has been dubbed the Functional Genomics Experiment model, or FuGE. The project is expected to serve as an overarching framework for data standards in genomics, transcriptomics, proteomics, metabolomics, and — hopefully — any other unforeseen "omics" that are yet to appear.
FuGE is essentially "future proof," according to James DeGreef, vice president of product management at GenoLogics. "We looked at a lot of different standards, and we like FuGE because they've put a lot of rigor into designing it well as a standard object model," he said.
The FuGE project is actively reaching beyond the microarray community to involve standards efforts in other research areas. Angel Pizarro, a FuGE developer and the director of the bioinformatics facility at the University of Pennsylvania's Institute for Translational Medicine and Therapeutics, gave a talk on FuGE during the Protein Standards Initiative's spring workshop in Siena, Italy, in April.
"If parties are working within the same data format, I think it will help determine what are the statistically significant results at the end of the day," Pizarro told BioInform's sister publication ProteoMonitor. "It will help, for example, RNA expression researchers and proteomics researchers to link together data from experiments in a systems biology way."
Development of FuGE began last September, when researchers from academic and commercial organizations — including the European Bioinformatics Institute, Stanford University, the University of Pennsylvania, the University of Glasgow, the University of Edinburgh, Rosetta Biosoftware, Affymetrix, GenoLogics, Applied Biosystems, Agilent Technologies, and Thermo Electron — began collaborating.
The FuGE developers met in April at the PSI workshop, and plan to meet again during the first week of August to wrap up the standard. By the first week of September, the group hopes to have a framework for producing XML schema that will give developers a real platform for developing FuGE-based software tools.
Pizarro described FuGE as more of a framework for developing standards than as a data standard in its own right. "The grand vision is to be able to provide this framework that people would be able to use as a standards model when developing software," he explained.
FuGE is simpler than MAGE, Pizarro said, and incorporates some of the "circular" workflows of the PEDRo standard, which accounts for proteomics' iterative experimental protocols.
Pizarro noted that FuGE will also include an ontology model that will be common across genomics, proteomics, metabolomics, and transcriptomics experiments.
"Ontologies, by their nature, are very system specific and can't be viewed outside of the system they were based in," said Pizarro. "What FuGE will do is reference external ontologies. It will query semantic validity in whatever system [the terminologies] were developed in."
Pizarro said that LIMS developers who deal with sample management should benefit most from FuGE, and GenoLogics' DeGreef agreed. He described FuGE as a framework for handling information related to laboratory management, sample tracking, and research workflow, while emerging proteomics file formats like mzData and mzXML are useful for the scientific data that "hangs off" that larger framework.
DeGreef said that GenoLogics plans to use FuGE in future versions of its ProteusLIMS software, but he did not provide a timeline for when that capability may become commercially available.
For now, he said, GenoLogics is collaborating with bioinformatics developers at the Fred Hutchinson Cancer Research Center on a markup language based on the FuGE object model called EXP-ML, for experiment markup language. The Hutch serves as the data repository for a 10-center biomarker discovery project, DeGreef said, and GenoLogics is partnering with the center to modify ProteusLIMS to support FuGE and EXP-ML so that all the data for the project can be imported and exported in those formats.
Further information on FuGE is available at http://sourceforge.net/projects/fuge/.