Evolutionary analyses across species have proven to be a powerful approach in elucidating the functional constraints of DNA elements, in particular protein-coding genes, but are less interpretable in the context of CRM architecture [6,7]. In part, this is due to the fact that CRMs often have a 'modular', rather than 'base-by-base', conservation that may escape detection by conventional alignment-based approaches [8]. Moreover, conservation in DNA binding profiles can be detected even without apparent DNA sequence constraint [9]. Even at the level of individual TFBSs, differences in sequence may be hard to interpret - as such differences, for example, may reflect evolutionary 'fine-tuning' to species-specific factors to preserve uniform outputs rather than signifying a lack of functional constraint [6,10-12].
