Pathways at MaizeGDB – Strategies for Curation and Data Sharing

Functional genomics became a reality for maize with the release of the draft sequence of 2.3 GB B73 inbred line in 2009. Gene expression datasets both at MaizeGDB and in other databases contain largely predicted functions. These rely on evidence compiled by UniProt and the GO Consortium, and where TAIR is the main contributor of higher plant data to these resources. We report here a new pathways curation effort at MaizeGDB and that leverages two distinct BioCyc Pathway-tools representations. One, CornCyc, is a very high stringency product generated by the Plant Metabolic Network project, which relies on the highly curated AraCyc.  The other, MaizeCyc, generated by the Gramene project, was less stringently computed and with greater likelihood for assigning a putative gene function to a gene model of interest. In addition to curation tools included in the Pathway Tools software suite, we have developed a Gene Ontology (GO) annotation module that supports community curation and provides robust validation via the NCBO portal using RESTful webservices.  Scripts have been developed that import GO annotations  from MaizeGDB to the object-oriented BioCyc databases.

We will initially curate pathways selected for relevance to important agronomic traits, availability of data in maize, especially mutants, and any suggestions from the MaizeGDB Working Group.  We are exploring several textmining approaches to enhance our potential for curation. Examples will be shown for some new tools, ”RLIMS-P” and “eFIP” developed at the PIR (Protein Information Network) to capture information about kinases and the interaction networks of phosphorylated proteins.