Exhaustive Discovery of Escherichia coli K-12 MG1655 Missing Genes Using Comparative Genomics Approaches

• Main Authors: Shan-Chun Yang, 楊善淳
• Other Authors: Chuan-Hsiung Chang
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/29992441771085539490

碩士 === 國立陽明大學 === 生物資訊研究所 === 94 === The numbers of sequenced bacterial genomes have been increasing rapidly over the past decade and many gene prediction programs like Glimmer, GeneMark and ZCurve were published. However, there are many genes may not be properly predicted by those programs. In thesis, we developed an ab initio method for bacterial gene prediction. We first obtained the candidate genes that are highly conserved in bacterial genomes except those in the E. coli. Because these highly conserved genes are assumed to be encoded as functional proteins through cross-species sequence comparison. Next, we filtered the candidate gene pool by removing overlapping genes and genes that do not have proper translation start site and having abnormal coding potential. In the result, there are 4,436 genes were predicted in E. coli K-12 MG1655. The sensitivity and specificity are 90.56% and 87.98% respectively. There are also 533 novel genes were reported by using our gene prediction method which are never found in other prediction methods. Eight of these novel genes are longer than 300bp and having ribosomal binding sites at the upstream regions. They are found to be conserved within two different species and were not overlapped with rRNA genes, tRNA genes or pseudogenes. In future work, wet lab experimental verification is required to identify the authenticity of these novel genes. This method can be applied to those currently available and upcoming bacterial genomes. The E. coli K12 gene prediction results are available on the web and can be accessed from: http://140.129.78.155/services/BGF/.