From profiles to function in epigenomics

A wealth of data is emerging from diverse studies of epigenomics, including genome-scale profiles of DNA methylation, histone modifications and higher-order chromatin features. In this Review, the authors discuss how, despite all this information, many challenges remain for inferring and proving the physiological and pathological functions of chromatin states. They describe the degrees of 'functionality' that are revealed by different experimental approaches, the value of integrative strategies and visions for the future. Myriads of epigenomic features have been comprehensively profiled in health and disease across cell types, tissues and individuals. Although current epigenomic approaches can infer function for chromatin marks through correlation, it remains challenging to establish which marks actually have causative roles in gene regulation and other processes. After revisiting how classical approaches have addressed this question in the past, we discuss the current state of epigenomic profiling and how functional information can be indirectly inferred. We also present new approaches that promise definitive functional answers, which are collectively referred to as 'epigenome editing'. In particular, we explore CRISPR-based technologies for single-locus and multi-locus manipulation. Finally, we discuss which level of function can be achieved with each approach and introduce emerging strategies for high-throughput progression from profiles to function.