So far we have been aggregating TFBSs position-by-position, which limits the scope of questions that could be addressed using these data. This has prompted us to devise a constraint metric that could be computed for individual motif instances and compared between heterologous TFBS subsets defined on the basis of their biological properties. The results presented above confirm the expected model that the deleterious effect of TFBS variation depends on how much it perturbs the motif consensus. Therefore, we proposed to express the deleterious effect of TFBS mutations in terms of 'mutational load', a known population genetics metric that combines the frequency of mutation with predicted phenotypic consequences that it causes [31,32] (see Materials and methods for details). We adapted this metric to use the reduction in PWM score associated with a mutation as a crude but computable measure of such phenotypic consequences. For example, the load of a motif instance for which no variation is observed equals zero, while the load of a motif instance with a common mutation mapping to it that results in a severe loss of PWM
