Genome evolution of fungal pathogens from the Magnaporthe oryzae/grisea clade.

Bgpi : Equipe 5Bgpi : Equipe 5The GEMO project aims at characterizing genomic determinants of pathogenicity and evolutionary events involved in adaptation of 9 isolates from the M. oryzae/grisea clade differing in their host specificity. Eight strains from M. oryzae species complex and one of the related species M. grisea have been sequenced and assembled. De novo structural gene annotation was carried out using Eugene (Schiex, 2001) to predict genes and REPET (Flutre, 2011) to annotate Transposable Elements (TEs) in these 9 genomes. Four of them exhibited large supplementary genomic regions potentially issued from an unknown bacterial strain of the Burkholderia genus. An original strategy based on Gotham software (Ménigaud, 2012) was used to accurately quantify these regions in all the affected genome scaffolds. Functional gene annotations were performed using InterProScan. Databases and interfaces relying on the GMOD tools (gmod.org) were set up to browse annotations and facilitate further evolutionary analyses. In order to identify gene families, the entire set of the predicted and known proteins of the M. oryzae/grisea genomes were clustered using OrthoMCL (Li, 2003). A total of 20443 clusters (15326 assigned to M.oryzae/grisea and 5117 to Burkholderia) were obtained, including 8154 clusters comprising single copy genes shared by all genomes (core genome) and variable number of genome specific gene families (305-1550). Genes encoding putative Secreted Proteins (SPs) were identified in 2522 OrthoMCL clusters (2271 of M. oryzae/grisea and 251 of Burkholderia). Further analyses regarding genome-specific and rice-infecting specific genes and SPs will be presented. OrthoMCL families were processed to infer the phylogenetic reference genealogy of the M. oryzae/grisae complex. We also evaluated the ability of individual genes to recover the same topology as that supported by most of the genes by using a recent method based on multiple co-inertia analysis (de Vienne, 2012). Most of the genes exhibit a concordant topology with the reference tree except a small set of 'outliers'. Further investigations are currently being performed to determine possible causes for incongruities. Finally, we present preliminary results regarding the comparison of TE distribution in M. oryzae/grisea species taking into account the reference genealogy of the strains