Systematic analysis of a biological system requires elucidation of its components. However, genome sequencing is only the first step; any analysis of transcription control and further functional genomics require the identification of all transcribed transcripts. FANTOM is the international consortium for the "functional annotation of mouse" or the "functional annotation of mammals", and produces data for analyzing the human and mouse transcriptome. To gain a substantial overview, FANTOM3, which is our latest milestone, has provided not only additional mouse full-length cDNAs, but also four datasets using new or familiar technologies; 1) 102,801 fully-sequenced and annotated cDNAs, derived from 237 full-length cDNA libraries; 2) 722,642 5-ESTs and 1,578,613 3-ESTs; 3) 7,151,511 mapped CAGE tags from 145 mouse libraries and 3,106,472 CAGE tags mapped from 24 human libraries, for identifying not only the precise start sites of transcription, but also relative expression and promoter usage; 4) 118,594 GIS and 968,201 GSC mapped ditags from four GIS and four GSC libraries for identifying the precise transcribed regions. Mapping of these transcripts drew the outline of the complex antisense transcription network on the genome, and many putative noncoding transcripts were shown to participate in such sense-antisense paring. Both sense and antisense transcripts tended to be co-expressed, and some transcripts influenced the copy number of the opposite transcript. Several previous studies using different cDNA synthesizing methods or microarray tiling experiments achieved the same results in other species, showing that the complex transcription and regulation by antisense transcription is a common feature in higher eukaryote genomes.
Genome Science Laboratory,Discovery and Research Institute, RIKEN Wako Main Campus, 2-1 Hirosawa,Wako, 351-0198, Japan.