Date: Tuesday, January 17, 2012
Time: 11:28 AM
Time: 11:28 AM
The best functional annotation for any gene is derived from published developmental, physiological or biochemical studies that directly examine the role of the gene in the biology of an organism. Unfortunately, extracting functional annotation from published literature is not practical for most genome annotation projects. Most functional annotation is assigned using an automated transitive process that is based in a large part on sequence and/or domain similarity. However, even after computational analysis, a large percentage of genes still lack a meaningful annotation. In the rice genome, 42% of all genes lack an assigned functional description. Using 15 publicly available expression data sets, we performed gene coexpression network analysis to identify modules of highly correlated genes to provide enhanced functional annotation to rice genes. These analyses have been performed on data from individual experiments to maintain the physiological relevance of the results. With these analyses, 13,537 genes have been assigned to 71 gene modules including 2,980 that lacked functional annotation (17% of all genes with no known function). Genes within these modules are now associated with groups of genes with similar expression patterns. The summation of the functional annotation of the genes in a single module provides a type of collective annotation for all genes in the module. Additionally, the expression pattern for each gene examined within a particular experiment provides a phenotypically based form of functional annotation that can be easily interpreted by a researcher. These data may be viewed at the MSU Rice Genome Annotation Project.