Regulatory Biophysics From Sequence Data

We demonstrate a new technique for probing the function of a regulatory sequence, using ultra-high-throughput sequencing to generate large data sets of the activity of mutant sequences. An information theoretic data analysis technique is then used to model the activity, avoiding assumptions about noise in the experiment. We apply this technique to the well studied \emph{lac} promoter in \emph{E coli.} and characterize the specificities of the DNA binding proteins in physical units and infer their \emph{in vivo} interaction energy.