The human genome reference assembly is crucial for aligning and analyzing sequence data, and for genome annotation, among other roles. However, the models and analysis assumptions that underlie the current assembly need revising to fully represent human sequence diversity. Improved analysis tools and updated data reporting formats are also required.