PhD Defence: Pernille Iversen

Reading Prokaryotic Genomes

Supervisor:
Professor Anders Krogh (KU)

Commitee:
Professor Kaj Frank Jensen (Chair, KU)
Regents' Professor Mark Borodovsky (Georgia Tech and Emory University)
Assoc. Professor Anders Gorm Pedersen (DTU)

Abstract:
Gene finding in prokaryotic organisisms is easier than in eukaryotes, and by many people it is thought to be a solved problem. However, previous studies have indicated that some genome annotations and many publicly available databases of protein information contain a large number of errors.
As part of the PhD work a large-scale study has been conducted, which shows that problems like annotation of too many short genes, wrong start codons, missing genes and overannotation still needs to be adressed. So correct localization and identification of prokaryotic genes is clearly not a solved problem.
Several computational gene finders exists, which can be valuable tools in genome annotation. For gene finders to be part of a larger annotation pipeline, they need to be automated and be able to adapt to the genome under consideration. In order for gene finders to do as good a job as possible, it is imperative that they are continuously refined. Further development of the genefinder EasyGene and subsequent utilization of it’s features in various projects has been a major part of the work for this thesis.