Selection of the appropriate target site is crucial to the success of an antisense experiment. The selection is difficult because RNAs fold to form secondary structures, rendering most of the molecule inaccessible to intermolecular base pairing with complementary nucleic acids. Conventional approaches, such as selection by 'sequence-walking' or computer-assisted design, have not brought significant success. Several empirical selection methods have been reported, a number of which are summarised in this review. Of notable significance are the 'global' methods based on mapping of transcripts with the endoribonuclease H (RNase H) and oligonucleotide scanning arrays.