Because of the increasing speed of computers, the TWINSCAN analysis of C. Elegans is able to use more accurate models of intron length than previous analyses. This is important for finding exons, which house the coding machinery of proteins. While getting intron length is helpful for gene annotation, the process is 15 times slower than the typical, less accurate methods. Being able to define intron length has implications for the human genome, which is much larger than C. elegans and has an average intron length of about 4,000 base pairs, compared with an average intron length of a couple hundred base pairs in C. elegans.
Brent and colleagues from the Dana-Farber Cancer Institute and Harvard Medical School published their findings in the April, 2005 issue of Genome Research. Brent's graduate student, Chaochun Wei, is first author on the paper. The research was supported by grants from NIH, NSF, the National Cancer Institute, the National Human Genome Research Institute, and the National Institute of General Medical Sciences.
Brent has brought his bioinformatics skills to many genomes, including those of mammals, other nematode species and most recently the fungus Cryptococcus neoformans. Brent's approach to gene prediction stands traditional genome annotation on its head because it starts with a computer analysis of the genome sequence, using that as a hypothesis designing experiments to test the hypothesis. The traditional modus operandi is a data-driven approach that starts with sequencing a random sample of tens of thousands of cDNA clones. Whereas the traditional approach leads to sequencing some genes thousands of times and others not at all, Brent's approach is to sequence each predicted gene once.
"I've been building a case that we should start with predictions," he said. "Each gene sequence is mor
Source:Washington University in St.Louis