Abstract
Aberrations in the regulatory genome play a pivotal role in population-level disease predisposition. Annotation of the regulatory regions using appropriate primary tissues – instead of cell lines affected by selection and other confounding factors – could shed new light into mechanisms underlying common conditions. We test this approach in uterine leiomyomas, highly prevalent benign neoplasms of the myometrium, by creating 15-state chromatin annotations for myometrium and uterine leiomyomas. Integration with RNA-seq, ATAC-seq, HiChIP and methylation data enables us to compare the epigenomes of myometrium and ULs with distinct driver mutations, highlighting the role of bivalent regions in the neoplastic process. Subsequently, a genome wide association study meta-analysis is performed, using three different cohorts. Disease association loci are enriched at active chromatin, especially at enhancers, and harbor tumor- and driver mutation-specific chromatin states. At SATB2 locus we show the effect of the risk genotype already in the normal tissue. Integration of genome-wide association studies and deep regulatory genomics data from the correct tissue type represents a powerful approach in understanding population-level disease predisposition.