Marc Hellmuth (Universität Greifswald)
Phylogenomics aims at finding plausible hypothesis about the evolutionary history of gene or species based on genomic sequence information. Genes are passed from generation to generation to the offspring. Some of those genes are frequently duplicated, mutate, or get lost - a mechanism that also ensures that new species can evolve. The history of a gene family is defined by a phylogenetic tree of the genes (the gene tree), together with an event labeling of its inner vertices and edges that identifies gene duplications, speciations events, and possibly horizontal gene transfer (the movement of genetic material between coexisting species).
Homology-relations are defined in terms of the edge- or vertex-labels of the ``true'' gene trees. However, the events of the past cannot be observed directly and hence, must be inferred from the genomic data available today. Intriguingly, it is possible in practice to distinguish certain types of homology-relations (e.g.\ orthologs, paralogs) using genomic sequence comparison without requiring any a priori knowledge on the topology of either the gene tree or the species
tree.
In this talk, we want to answer the questions:
(1) How much phylogenetic signal is contained in the homology-relations? In other words, is it possible to infer phylogenetic trees from those relations?
(2) Can we characterize homology-relations?
To this end, we study the graph structure of the those relation and provide efficient algorithms to recognize mathematically correct homology-relations and to reconstruct the underlying phylogenetic trees.
