Tagging mammalian transcription complexity

The nature of the ‘transcriptome’ is more complex than first realized. Although CAGE, various tagging technologies and tiling arrays show that most of the mammalian genome is transcribed, a large proportion of transcripts do not encode proteins and are either poorly polyadenylated, involved in sense–antisense pairs or never leave the nucleus. In this article, I review the various techniques and data sets that are currently used to measure gene transcription and the evidence that reveals the true extent of transcription in mammalian genomes. The next few years will see efforts to identify novel transcripts systematically and decipher their function. A deeper understanding of transcriptional complexity might even lead us to redefine what we mean by the term ‘gene’.