Given a sequence database, we would like to find all

Given a sequence database, we would like to find all

Given a sequence database, we would like to find all sequential patterns that do not only satisfy a minimum supportmin_sup, but also start with “{a}” and end with “{b}”, where {a} or {b} is an itemset containing a single item a or b, respectively. How would you modify the PrefixSpan algorithm so that such patterns can be generated efficiently?