As a stripped-down baseline, we applied PhyloScan in a scan of the full E. coli sequence database, ignoring all other sequence data; this baseline is equivalent to the original Staden method, and thus has the same statistical power.