Files in this data supplement:
- Supporting Information: Figures S1-S9 (PDF, 1.6 MB)
- Figure S1: The average information content per nucleotide of binding sites range from around 0.25 bits to 2 bits, in both eukaryotes (top, 454 curated transcription factor motifs) and prokaryotes (bottom, 79 motifs) (PDF, 137 KB)
- Figure S2: When binding site length n is too small (compared to the stable region), selection to increase n is strong (top) (PDF, 163 KB)
- Figure S3: Varying model parameters by several orders of magnitude does not alter our qualitative results on the stable region of binding site length and information content, or its agreement with empirical data (PDF, 174 KB)
- Figure S4: Binding sites with lengths that start above (light gray line) or below (dark gray line) the stable region (dashed black lines) both evolve towards values inside the stable region (PDF, 177 KB)
- Figure S5: (Top) Sample paths for evolving binding site length and degeneracy. Populations started outside of the stable region evolve to values inside it. (Bottom) The distribution of n and r for 102 binding sites after n and r have been allowed to evolve to equilibrium. (PDF, 186 KB)
- Figure S6: Sexual reproduction increases the range of stable binding lengths (PDF, 134 KB)
- Figure S7: The more complex biophysical model produces stable regions that show good qualitative agreement with those produced by the more simple model (PDF, 272 KB)
- Figure S8: In the more complex model, just as in the more simple model, when binding site length n is short (compared to the stable region), selection to increase n is strong (top) (PDF, 174 KB)
- Figure S9: The distribution of n and r for 102 binding sites after n and r have been allowed to evolve to equilibrium (PDF, 153 KB)