I just wondered about the maxScore calculation for PWMs of different types ('log2probratio' vs 'prob' as given in the help page of PWM).
I would have assumed that for 'log2probratio' the maxScore is the sum of the column-wise maximum of the PWM. This also holds. In contrast, for 'prob' PWM I would have assumed that the maxScore is the product of the column-wise Maximum of the PWM (since it is not on log-scale).
However it seems that for both 'log2probratio' and 'prob' PWMs the maxScore is always the sum of the column-wise maximum of the PWM.
What am I getting wrong?