My problem is when I plot the same data with linear and with log-transformed axes, the equation that is calculated by stat_poly_eq gives is different between both plots (wrong in the log-transformed plot). What could be the reason for that, and is there a simple way to fix that?