In both cases, 95% confidence intervals were calculated by interpreting the fractions of counts as binomial rates and then using Wilson’s method for binomial confidence intervals as implemented in the R package binom (49).