Overlapping Antisense Transcription in the Human Genome
Accumulating evidence indicates an important role for non-coding RNA molecules in eukaryotic cell regulation. A small number of coding and non-coding overlapping antisense transcripts (OATs) in eukaryotes have been reported, some of which regulate expression of the corresponding sense transcript. The prevalence of this phenomenon is unknown, but there may be an enrichment of such transcripts at imprinted gene loci. Taking a bioinformatics approach, we systematically searched a human mRNA database (RefSeq) for complementary regions that might facilitate pairing with other transcripts. We report 56 pairs of overlapping transcripts, in which each member of the pair is transcribed from the same locus. This allows us to make an estimate of 1000 for the minimum number of such transcript pairs in the entire human genome. This is a surprisingly large number of overlapping gene pairs and, clearly, some of the overlaps may not be functionally significant. Nonetheless, this may indicate an important general role for overlapping antisense control in gene regulation. EST databases were also investigated in order to address the prevalence of cases of imprinted genes with associated non-coding overlapping, antisense transcripts. However, EST databases were found to be completely inappropriate for this purpose.