QTL Mapping and Genomic Prediction of Disease Reactions of Rice to the Rice Blast Fungus M. oryzae Races

Blast disease of rice caused by the fungus Magnaporthe oryzae is one of the most damaging diseases of rice worldwide. Major blast resistance genes have been deployed for broad-spectrum resistance. However, the minor genes and epistatic interactions contributing to the broad-spectrum resistance are largely uncharacterized. Using a mapping population of 272 individuals derived from the blast resistant variety ‘Katy’ and blast susceptible ‘PI312777’, we identified and verified resistance spectra of major genes Pi-ta/Ptr and Pi-ks and further identified one additional major and 13 minor resistance quantitative trait loci (QTLs) for 10 of the 11 pathogen races tested and detected additive-by-additive epistasis for 32 pairs of QTLs. We used Bayesian network (BN) structure learning to model associations between QTLs, blast resistance, and agronomic traits, revealing four traits associated with resistance to specific pathogen races. The major gene cluster Pita/Ptr conferred resistance to only 7 of 11 races, highlighting the need for multiple smaller effect resistance QTL to achieve broad-spectrum resistance. Genomic predictive abilities (GPAs) were calculated and were lowest for resistance to race IE1k (for which Pita/Ptr is not effective) and highest for IA45. Reciprocal GPAs were calculated for prediction across pairs of blast races and were lowest for pathogen races IB1, ID1 and IE1k, moderate for races IB49, IE1, IB54, IH1 and highest among IC17, IG1, IA45, IB45. A genomic selection rank sum index (GSRI) of genomic estimated breeding values (GEBVs) for overall resistance to 11 pathogen races distinctly separated high and low resistant lines for all races except IE1k.