GALLUT, C; Universit� Pierre et Marie Curie: Parsimony inference from genome organisation

Phylogenetic parsimony analyses rely on homologous characters comparison, there by identifying homologies is an inescapable first step, the second step is to code these homologies in a suitable way to perform the analyse. It is very much the same for phylogenetic analyses based on genome organisation. Classically genomes are compared at nucleotidic sequence level for phylogeny reconstruction. The sequences of all the taxa in the analysis are aligned and each nucleotid position resulting of this alignment is then considered as a homologous character. Implicitly this reasoning infers the phylogeny by only considering the evolutionary events at sequence level. But genomes bear phylogenetic information at higher level of organisation than the sequence level. In this regard we have considered different possible genome organisations and proposed different coding strategies to inferring phylogenetic relationships. Genomes are made of one or more chromosomes that are whether circular or linear. In our view a chromosome is an ordered sequence of homologous components like genes, clusters of genes or on higher range, identical sets of G bands on karyotypes. The chromosome topology is represented by three different features, the presence or absence of homologous components, their relative position to the others and if possible their orientation. This strategy is free of a priori hypotheses about evolutionary events that could have occurred between the taxa under scrutiny but upon the resulting tree(s) it is usually possible to reconstruct the genome topology of hypothetical ancestors thus allowing to inferring the evolutionary processes of the genomes. In this, our approach differs from other gene order phylogenetic methods that most generally seek for the minimum evolutionary events like inversions, transpositions or translocations. We both illustrate our method with theoretical and biological examples.