Coordinate Systems for Supergenomes

Background: Genome sequences and genome annotation data have become available at ever increasing rates in response to the rapid progress in sequencing technologies. As a consequence the demand for methods supporting comparative, evolutionary analysis is also growing. In particular, efficient tools to visualize -omics data simultaneously for multiple species are sorely lacking. A first and crucial step in this direction is the construction of a common coordinate system. Since genomes not only differ by rearrangements but also by large insertions, deletions, and duplications, the use of a single reference genome is insufficient, in particular when the number of species becomes large. Results: The computational problem then becomes to determine an order and orientations of optimal local alignments that are as co-linear as possible with all the genome sequences. We show that this problem can be phrased as a particular variant of the Betweenness Problem that is NP hard in general. As exact solutions are beyond reach for the problem sizes of practical interest, we introduce a collection of heuristic simplifiers to resolve ordering conflicts. Conclusion: Benchmarks on real-life data ranging from bacterial to fly genomes demonstrate the feasibility of computing good common coordinate systems.