This method can similarly be applied to several types of transcriptional regulatory elements including promoters and repressors, and can be used to uncover the underlying architecture of these elements as well as to learn the effect size distribution of mutations in regulatory regions.