Large-scale tests of the algorithm, using sequence data from a selection of 1,820 human primary transcripts, demonstrated that the inclusion of ESE and ESS motif data considerably improved the ...