Fig. 2.

An illustration of growing a pattern tree using the bounded-prefix-growth procedure. A node in the pattern tree denotes a pattern candidate, accompanied by its complete projected database. The projected database is the collection of sequences with a match to the qualified sequence suffix pattern in the forward or reverse complement DNA sequence. The support for a pattern is defined as the ratio of distinct sequences (S1, S2, and so on), each with at least one match to the pattern over the total number of sequences in the original database. A pattern may match a sequence at more than one position. These sequences are denoted as S1.1, S1.2, and S1.3 in the projected database.