3.5 | Gene content in the candidate genomic regions
A total of 118 genes distributed in 25/32 candidate regions were identified using the annotated reference genome of P. fijiensis(Arango Isaza et al., 2016). However, due to linkage disequilibrium and the hitchhiking effects, it is unlikely that all these genes are involved in pathogenicity and adaptation. The number of genes per region ranged from 1 to 12. All the information on these genes available to date is provided in supplementary (Table S5). None of the candidate genes corresponded to the putative small secreted proteins (SSPs) identifiedin silico in Arango-Isaza et al. 2016, which are putative effectors involved in plant-pathogen interactions. However, secreted molecules other than SSPs can also be effectors (Rovenich et al., 2014). We identified 10 genes distributed across nine regions with a peptide signal that could correspond to effectors (Table 5). Only one gene among these genes was identified in the in-vitro secretome of P. fijiensis published by Escobar-Tovar et al. 2015. Four other genes found in different regions were associated with increased expression in leaf tissues infected by P. fijiensis in the transcriptome study published by Noar & Daub 2016. Finally, we found 17 genes distributed in 10 genomic regions corresponding to homologous genes indexed in the PHI-base (a database containing more than 6 000 genes involved in host-pathogen interactions, (Urban et al., 2015)) that are known to lead to a total loss, reduction, or gain in pathogenicity in other fungal species. A total of 12 genes located in the 17 regions significantly correlated with DLA-R and/or DLA-S were also identified in at least one of the other previous analyses. Based on all the data available at present, we considered several genes as candidates involved in quantitative pathogenicity and host adaptation (highlighted in Table 5 and details in Table S5): three major candidate genes (highlighted in red in Table 5 and Table S5) were located in regions significantly correlated with DLA-R or DLA-S and previously identified in at least two of the databases consulted, other good candidates (highlighted in grey) were located in regions correlated to DLA-S or DLA-R and identified in one of the database (9 genes) or found in at least two of the databases consulted (4 genes).