Two technology revolutions in genome sciences, next generation sequencing and genome-editing, have made it possible to start to understand the relationships between genotype and phenotype in molecular terms and herald an era of predictive biology. For non-coding control elements in the genome, which are the major targets of variation within and between species, analysis of the function of sequence variation is not straightforward. We must first identify those elements of the non-coding DNA that are functional and sensitive to variation. I will first introduce some of the recent work of the FANTOM Consortium, which has contributed to functional annotation of the human and mouse genomes. I will then talk about progress in the use of transcriptomic data to produce an expression atlas for each of the major livestock species and the ways in which coexpression clustering can suggest functions of genes that currently lack annotation. Finally, I will discuss the applications of genomic technologies to produce sustainable improvements in livestock productivity and the use of large animals as model organisms.