The Human Proteome Organization’s Protein Standards Initiative held a workshop in Carros, France, last week, where participants in three different tracks worked to define the next phase of the PSI’s molecular interaction standards, lay out plans for finalizing the first version of its M/Z mass spectrometry standard, and develop definitive standards for the submission of proteomics experimental data to journals.
“Hopefully we will in the end unite all of these aspects and have one common model — but that’s still quite awhile away,” Henning Hermjakob, head of the PSI Molecular Interactions track, told BioInform’s sister publication ProteoMonitor.
Leading the way among the three groups, the PSI Molecular Interactions effort published a paper on PSI-MI, its XML-based standardized format for molecular interactions, in Nature Biotechnology in February. Hermjakob, sequence database group coordinator at the European Bioinformatics Institute, said that the next release of PSI-MI is planned for September, just before the HUPO meeting in Beijing. Major changes from the first version include an expansion of interaction data to include protein interactions with DNA, RNA, and small molecules, as well as an emphasis on using either UniProt identifiers or NCBI GI numbers for every protein in the file “to make it easier to detect redundancies between different data sets,” Hermjakob said. In addition, BIND and MIPS have agreed to release their data in the new format.
The mass spec data standards track, led by Weimin Zhu, head of database applications at EBI, plans to finalize changes to the M/Z data format within the next month, and has made a deal with “one or two of the vendors … [to] implement the M/Z data format as their native output format, and we will then import this directly into the search engine,” Hermjakob told ProteoMonitor. He would not say which vendors have agreed to do this, but did say they were among the industry representatives at the workshop — which included Bruker Biosciences, Shimadzu, MDS Sciex, and Thermo Electron. Hermjakob said that more information about this collaboration would be released at the ASMS meeting in Nashville next month.
The mass spec group also began to develop a search engine output format, Hermjakob said, which will be available in a preliminary version for HUPO Beijing.
The experimental data submission effort, led by Chris Taylor of the University of Manchester, is at the earliest stage in its development, but it has now officially adopted MIAPE — or minimum information about a proteomics experiment — as its model. MIAPE is based on the MIAME guidelines for microarray data. Analogous to the MAGE-OM object model and MAGE-ML markup language that correspond to MIAME, plans are in the works to create PSI-OM and PSI-ML for proteomics experiments. Hermjakob said that preliminary MIAPE results should be ready for HUPO Beijing in October.
HUPO PSI will hold its next major workshop in a year.