Polymorphism Detection by Microsatellite Markers in a Magnaporthe oryzae Population From Different Geographical Areas of Brazil

This study aimed to examine Brazilian M. oryzae populations using 18 microsatellites. Fifty cultivars were sown in plastic trays for the pathotyping of 847 isolates. The DNA of 494 isolates was extracted and purified using the modified Doyle and Doyle method, the genetic structure was determined by the software Structure, and the actual number was selected from the prediction method based on the K values. Nei's genetic distance among the subpopulations was determined with the aid of the program Genetix, and the amova was performed with the program Arlequin. Out of 847 inoculated monosporic isolates, 528 infected their respective cultivars; of the 528 isolates pathotyped, there was a prevalence of group IA and pathotype IF-1, which was the most frequent pathotype in the rice production areas of Brazil. The Bayesian clustering analysis indicated that 19 was the optimal value of K; this value was the lowest standard deviation and log (ln K) closest to zero, which predicted the 494 isolates of M. oryzae that were selected for molecular studies to be grouped into 19 subpopulations. The AMOVA detected a 37.13% variability within the 19 subpopulations and 62.87% variability among the subpopulations. The polymorphic information content (PIC) ranged from 0 to 0.756. Thirty three rare alleles were found distributed among 15 out of 19 subpopulations. The Margalef index ranged from 38.69 to 79.21 for all 18 analysed locus. The results indicated that the identification of different blast resistance genes must consider the composition of each subpopulation and that the identification is most effective when performed within a subpopulation and then between subpopulations.