Analyzing the genome of Halorhabdus utahensis using three different databases

Samantha Simpson, Peter Bakke, Nick Carney, Will DeLoache, Mary Gearing, Matt Lotz, Jay McNair, Pallavi Penumetcha, Laura Voss, Max Win, Laurie J. Heyer, A. Malcolm Campbell Department of Biology, Davidson College Department of Mathematics, Davidson College Abstract One way to harvest information about a species is to automatically annotate its genome using a specialized computer tool, and then curate the genome manually based on the tool’s annotation. Several annotation websites exist that accept sequenced genomes, but they are not all the same. The Halorhabdus utahensis genome was submitted to three tools to be automatically annotated: RAST, TIGR’s Manatee server, and JGI’s Integrated Microbial Genomes’ (IMG) site. The three servers predicted different genes due to internal biases and theories governing the creation of the annotation tool. RAST, for example, is more likely to predict longer genes that use alternative start codons (not ATG). The differences in annotation tools highlight genes of interest – genes predicted by only one or two of the databases – that require further study. After annotating H. utahensis, pathway analyses allow the user to infer function based on genes that are predicted in the genome.