Definition 4 (Pattern probability)
Given a pattern P = P1, P2, ..., Pq (Pi is a DNA alphabet) and a DNA sequence database D, the pattern probability of P in D is defined as , where Pr(Pi, D) = # of occurrences of an alphabet Pi/|D|.
For example, the pattern probability of pattern "ATCG" in Table 1 is Pr(ATCG,D) = Pr(A,D) × Pr(T,D) × Pr(C,D) × Pr(G,D) = (10/55) × (18/55) × (12/55) × (15/55) = 0.182 × 0.372 × 0.218 × 0.273 = 0.00403.