is absent in the other, was not considered in this analysis because of the difficulty in distinguishing true intron retention events from contamination of the EST databases by pre-mRNA or genomic sequences. The presence of these and other artifacts in EST databases are important caveats to any analysis of EST sequence data. Therefore, we imposed stringent filters on the quality of EST to genomic alignments used in this analysis, accepting only about one-fifth of all EST alignments obtained (see Materials and methods).
