 | Fine Structural Analysis of the Transcription Start Sites of Human mRNA
|
| | Yutaka Suzuki1)3), Hirotoshi Taira4), Tatsuhiko Tsunoda2), Junko Mizushima-Sugano1), Jun Sese2), Hiroko Hata1), Toshio Ota5), Takao Isogai5), Yoshiyuki Sasaki2)3), Toshihiro Tanaka2), Yusuke Nakamura2), Shinichi Morishita2), Kousaku Okubo6), Akira Suyama7) and Sumio Sugano1) |
| 1) Department of Virology, Institute of Medical Science, the University of Tokyo
2) Genome Center, Institute of Medical Science, the Unviersity of Tokyo
3) Genome Science Center, Institute of Physical and Chemical Research (RIKEN)
4) Intelligent Communication Laboratory, NTT Communication Science Laboratories
5) Helix Research Institute
6) Institute of Molecular and Cell Biology, the University of Osaka
7) Department of Life Sciences, the University of Tokyo |
| | One-pass sequencing of 100, 000 clones from cDNA libraries constructed by “Oligo-capping” enabled us to accumulate 5′-end sequences of 2251 kinds of named genes. As for redundantly isolated genes, we compared the 5′-ends of “Oligo-capped” cDNAs with each other. Unexpectedly the exact 5′-ends were heterogeneous in most cases. We selected 276 genes whose mRNA start sites were represented by more than five “Oligo-capped” cDNAs and mapped 6308 mRNA start sites in total onto the corresponding genomic sequences. Statistical analysis on these mRNA start sites revealed that the start sites were distributed over 61.7bp with the standard deviation of 19.5 on average. Statistical learning using a learner decision tree algorithm C4.5 revealed that the most discriminative feature of the genes with tightly clustered transcription start sites was the presence of TATA box in the promoter. It was also shown that the Oct-1 and CAAT box have synergistic and inhibitory effect on the presence of TATA box, respectively. |
| | Key Words: | Transcription start site, full-length cDNA, Oligo-capping |
| |
