Validation of genome-wide association studies as a tool to identify virulence factors in parastagonospora nodorum

Abstract

Parastagonospora nodorum is a necrotrophic fungal pathogen causing Septoria nodorum blotch on wheat. We have identified nine necrotrophic effector–host dominant sensitivity gene interactions, and we have cloned three of the necrotrophic effector genes, including SnToxA, SnTox1, and SnTox3. Because sexual populations of P. nodorum are difficult to develop under lab conditions, genome-wide association study (GWAS) is the best population genomic approach to identify genomic regions associated with traits using natural populations. In this article, we used a global collection of 191 P. nodorum isolates from which we identified 2,983 single-nucleotide polymorphism (SNP) markers and gene markers for SnToxA and SnTox3 to evaluate the power of GWAS on two popular wheat breeding lines that were sensitive to SnToxA and SnTox3. Strong marker trait associations (MTA) with P. nodorum virulence that mapped to SnTox3 and SnToxA were first identified using the marker set described above. A novel locus in the P. nodorum genome associated with virulence was also identified as a result of this analysis. To evaluate whether a sufficient level of marker saturation was available, we designed a set of primers every 1 kb in the genomic regions containing SnToxA and SnTox3. Polymerase chain reaction amplification was performed across the 191 isolates and the presence/absence polymorphism was scored and used as the genotype. The marker proximity necessary to identify MTA flanking SnToxA and SnTox3 ranged from 4 to 5 and 1 to 7 kb, respectively.Similar analysis was performed on the novel locus. Using a 45% missing data threshold, two more SNP were identified spanning a 4.6-kb genomic region at the novel locus. These results showed that the rate of linkage disequilibrium (LD) decay in P. nodorum and, likely, other fungi is high compared with plants and animals. The fast LD decay in P. nodorum is an advantage only if sufficient marker density is attained. Based on our results with the SnToxA and SnTox3 regions, markers are needed every 9 or 8 kb, respectively, or in every gene, to guarantee that genes associated with a quantitative trait such as virulence are not missed.