Using a comparative computational approach, Yáñez-Cuna et al. studied thousands of enhancers in three Drosophila melanogaster cell lines and determined sequence features of enhancers, including enrichment of dinucleotide repeat sequences. Experimental validation of these dinucleotide repeats showed that they are sufficient to form de novo enhancers from non-functional sequences. These results have implications for the identification of novel enhancers and for genome sequencing annotations in general.