Exploiting Gene Clusters to Curate Annotations

Wednesday, October 22, 2003 - 3:00pm - 3:50pm
Keller 3-180
Ross Overbeek (Argonne National Laboratory)
Previously, we argued that gene clustering on prokaryotic genomes was the key to locating missing genes, and we demonstrated that the technique worked remarkably well. The use of clusters is also the key to straightening out many of the assignments that could not be made precisely based only on similarities and motifs. We will consider the case of gene clusters related to leucine degradation as an example; they occur in phylogentically diverse organisms, and many of the genes involved currently have inaccurate or imprecise annotations. Comparative analysis of clusters, as well as occurrence profiles, can be used to methodically construct chains of assignments that follow from a few basic observations. This sets the stage where a single carefully chosen wet lab confirmation can confirm or reject a large number of assignments, often removing ambiguities from tens if not hundreds of genes.