Annotation of bovine, human and other relevant genomes

Leader: Chris Elsik (University of Missouri)

Task 8.1. Updated gene annotation using SNP and expression data
Genomic data collected above will identify genomic locations associated with traits of interest, under selection, or showing changes in expression or epigenetic regulation. We propose a feed-back annotation scheme to integrate these new data into the genome annotation, both as gene-level metadata, as well a indications of locations where annotation is incomplete or incorrect.

Task 8.2. Orthology, function and evolution
We will use the concept of orthology to allow bidirectional functional inferences between the bovine genome and model organisms such as humans and mice. Thus, we will link the association mapping results from WP4 & WP5 not only to the bovine annotation, but also to the human and mouse genome using our orthology inference package. Equally importantly, we will take well-developed network biology resources from mice and humans and integrate them with the data and predictive approaches from WP2 &WP4.

Task 8.2A. Automated orthology prediction for mammalian genomes and annotation linkage.
Our published orthology inference method (Bekaert and Conant, 2011) uses gene order data (synteny) in cases where sequence analysis alone cannot guarantee orthology. Once we have a set of human-murine-bovine orthologs, we can link the genomic association data from WP4 and WP5 to the human and mouse genomes, allowing both new functional information for human to be inferred from the bovine studies and allowing known human or murine functional information to be applied to bovine systems.

Task 8.2B. A bovine metabolic network.
We have already mapped the human metabolic network (a nearly complete catalog of human metabolic reactions and intermediary compounds) onto the bovine genome. After improving this mapping with new data, we will use flux-balance analysis to infer fluxes for all human/bovine enzymes and provide a link between the metabolic markers from WP2 & WP3 and the genomic variants associated with such changes (WP4 & WP5). This approach will allow us to infer enzymes whose expression level should influence the concentrations of the metabolic markers chosen in WP2 & WP3. In addition to allowing validation of genomic associations, these links will provide biological context for those linkage studies.

We will also link data from genome association studies in cattle to mutant phenotypes in mice and rats, disease phenotypes in humans to identify subnetworks of genes associated with phenotypes. This orthology and network-based approach can identify critical genes that are missed by association studies because no alleles of strong affect are segregating.

Task 8.2C. An integrated database solution for maintaining network and orthology inferences.
We will maintain the above network structures, inferences and orthology data in the Bovine Genome Database (BGD). We will both use NCBI’s Gene Linkout feature to link human and mouse genes back to bovine orthologs and create a Biomedical-to-Bovine feature in BGD that will encourage biomedical researchers to explore how bovine genomics can provide insight into human health. Traditionally, the bovine genomics community has relied using discoveries in humans and mice to make inferences about the bovine genome. With new efficient genomic technologies, we are now in the position to turn bovine into a model to transfer discovery back to humans. Thus, we will bring to the genomic community a strongly comparative approach, emphasizing the sharing of data and annotation across all mammalian genomes, including from human to bovine and back again.

WP2 WP3 WP4 | WP5 | WP6 | WP7 WP8 WP9