where Ei(m|p0) is the expected number of motif m's occurrences in sequence i assuming the background model p0, and L is the total number of sequences in the dataset.