Perhaps the greatest surprise of genetic studies of human disease is that 90% of top-scoring disease-associated loci lie outside protein-coding regions. This has increased the urgency of mapping non-coding DNA elements and regulatory circuits, in order to understand the molecular basis of human disease. To address this challenge, the Roadmap Epigenomics program has sought to systematically characterize the epigenomic landscape in diverse primary human cells and tissues, resulting in the annotation of 2.3M enhancer elements across 127 tissues and cell types, and tissue-specific regulatory networks linking enhancers to their upstream regulators and target genes. In this talk, I will describe the use of these annotations for understanding the molecular basis of genetic differences in common disease and cancer: (1) We uncover the mechanistic basis of GWAS hits, predicting and experimentally validating the causal variants, the cell types in which they act, the upstream regulators that target them, the downstream genes they target, and their molecular, cellular and organismal phenotypes in the context of obesity. (2) We combine genetic and epigenomic evidence to prioritize and experimentally validate weakly-associated variants in the context of cardiac repolarization phenotypes, showing that epigenomic data enables robust discovery with much smaller cohorts. (3) We use our regulatory predictions to identify new cancer genes based on recurrent somatic mutations in their linked upstream regulatory elements, revealing out-of-context de-repression as a common cancer strategy in the context of prostate cancer. These three applications, spanning the spectrum of common, rare, and somatic variants, illustrate the power and broad applicability of epigenomic annotations and regulatory networks for understanding human disease and cancer.