Feature vector encoded for a miRNA:target duplex. Each duplex is partitioned into two parts, with the first part (the seed) covers nucleotide 1 to 9 from the 5' end of the miRNA, and the second part (the non-seed) covers the rest of the duplex. Five types of base-pairing are defined. A total of 465 features, consisting of the 1-gram, 2-gram and 3-gram frequencies of the five pairing types, are encoded into a vector representing a miRNA:target duplex.
Yang et al. BMC Bioinformatics 2008 9(Suppl 12):S4 doi:10.1186/1471-2105-9-S12-S4