Comparative Genomics Integrated with Association Analysis Identifies Candidate Effector Genes Corresponding to in Phenotype-Paired Isolates from Australia.

Leaf rust is one of the most common and damaging diseases of wheat, and is caused by an obligate biotrophic basidiomycete, (). In the present study, 20 isolates from Australia, comprising 10 phenotype-matched pairs with contrasting pathogenicity for , were analyzed using whole genome sequencing. Compared to the reference genome of the American isolate 1-1 BBBD Race 1, an average of 404,690 single nucleotide polymorphisms (SNPs) per isolate was found and the proportion of heterozygous SNPs was above 87% in the majority of the isolates, demonstrating a high level of polymorphism and a high rate of heterozygosity. From the genome-wide SNPs, a phylogenetic tree was inferred, which consisted of a large clade of 15 isolates representing diverse presumed clonal lineages including 14 closely related isolates and the more diverged isolate 670028, and a small clade of five isolates characterized by lower heterozygosity level. Principle component analysis detected three distinct clusters, corresponding exactly to the two major subsets of the small clade and the large clade comprising all 15 isolates without further separation of isolate 670028. While genome-wide association analysis identified 302 genes harboring at least one SNP associated with virulence ( < 0.05), a Wilcoxon rank sum test revealed that 36 and 68 genes had significant ( < 0.05) and marginally significant ( < 0.1) differences in the counts of non-synonymous mutations between avirulent and virulent groups, respectively. Twenty of these genes were predicted to have a signal peptide without a transmembrane segment, and hence identified as candidate effector genes corresponding to . SNP analysis also implicated the potential involvement of epigenetics and small RNA in pathogenicity. Future studies are thus warranted to investigate the biological functions of the candidate effectors as well as the gene regulation mechanisms at epigenetic and post-transcription levels. Our study is the first to integrate phenotype-genotype association with effector prediction in genomes, an approach that may circumvent some of the technical difficulties in working with obligate rust fungi and accelerate avirulence gene identification.