Co-transcription of genes into single transcripts: another regulatory mechanism for gene expression in vertebrates

Todd D Taylor, Tulika P Srivastava, Vineet K Sharma, Yuichiro Nishida, Takayoshi Fujikake, Tadayuki Takeda, Ritsuko Ozawa, Maki Mushiake, Reina Okumura, Yoshiyuki Sakaki
RIKEN Genomic Sciences Center, 1-7-22 Suehiro-cho, Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan

Co-transcription of two distinct genes (child genes) into a single transcript (conjoined gene) has not been well explored. Either it is a somewhat rare phenomenon, or the current methods of genome annotation are not sensitive enough to identify such genes. Towards this we have designed a new computational algorithm "Conjoin" for the identification of conjoined genes in any genome given its messenger RNA or EST information. Applying Conjoin to the human genome, we have identified nearly 900 conjoined genes of variable lengths, some with multiple isoforms. We have so far confirmed the existence of 195 of these conjoined genes using RT-PCR and sequencing. In view of the fact that we observed several cases of conjoined genes occurring in the human genome, it appears that these conjoined transcripts are arising out of novel functional requirements and are not merely artifacts of transcription. However, the underlying mechanism controlling the formation of such conjoined genes in human and other vertebrate genomes remains to be explored. In order to confirm the presence of conjoined genes in other vertebrates, we implemented the Conjoin algorithm on both the mouse and chimpanzee genomes. It is remarkable to observe that the number of conjoined genes in mouse is far less than that in human, even though there is roughly the same amount of mRNA/EST data available. Thus it appears that the conjoined genes might be performing some novel functions and are contributing to human complexity as compared to other lower organisms. Therefore, we carried out a detailed functional analysis of the human conjoined and participating child genes. Further, in order to explore the intrinsic mechanisms of this process, the 5' and 3' flanking regions of the child genes were analyzed to search for the presence of any alternate or common regulatory elements that might be controlling the formation of conjoined genes.