Prof Dr Mario Stanke is currently Chair of Bioinformatics at the Institute of Mathematics and Informatics at the University of Greifswald, Germany, and gives a seminar as part of the interview process for a position as group leader at CBU and the Department of Informatics.
Title of talk:
“The sequencing of clades has become common, e.g. in the ongoing Genome 10K and Bird 10K projects, where the sequenced genomes are close enough to have widely conserved gene structures and yet are diverse enough so the differences between genomes shed light on selection. A large part of the talk will be about our recent comparative gene prediction methods on the basis of our gene-finder AUGUSTUS. The problem of finding intron-exon gene structures for all genomes of such a clade can be stated as one of finding a set of paths in a graph, thereby maximizing a score that includes comparative features such as selection, conservation and evolutionary changes to gene structures. Although maximizing this score is NP-hard, in many cases an exact solution can still be found using a dual decomposition approximation. This is akin to maximum a posteriori (MAP) inference on probabilistic graphical models where a large, intractable problem is divided into independently solved ‘slave’ problems. Applying the new comparative AUGUSTUS to 20 newly sequenced mouse strains or species, we could identify a novel, conserved protein-coding gene with even 184 exons.”
More info about the speaker here.