Gametic phase disequilibria in populations of race 2 and race 3 of Cochliobolus carbonum

Over 3 years 15 samples of C. carbonum were collected from 11 corn fields in North Carolina and Tennessee. Among 514 isolates of race 2 and 319 isolates of race 3, five phenotypic characters (mating type, production of pseudothecia, production of asci and ascospores, tolerance to cycloheximide and carboxin, respectively) that are controlled by single genes at unlinked loci (Mat, Psu, Asc, Cyh, Crb) were examined. Gametic phase disequilibrium (GPD) was analyzed by three methods. First, observed and expected four-locus haplotype frequencies were compared in G-tests for goodness of fit (method 1) and second, four-locus disequilibrium was calculated using an index of association which is based on the variance of the number of loci at which two isolates in a given population differ (method 2). Third, observed and expected frequencies of allele pairs were compared in 2×2 contingency tables (method 3). Each test was performed on individual and pooled samples which also allowed assessment of heterogeneity among samples. In general this heterogeneity was low, i.e. the consistency of associations among samples was high. Four-locus disequilibrium was significant in race 2 with method 1 but not with method 2. In race 3 we found no disequilibrium with any of the two methods. Method 3 indicated that several allele pairs were significantly associated in race 2 but not race 3. Thus GPD was significant in race 2 but not in race 3. Mating type frequencies were close to 0.50 in race 2 and race 3. This and the non-significant index of association indicate that the sexual stage of C. carbonum is operating within each of the two race populations. It is suggested that selection favours different haplotypes in the asexual than in the sexual stage of the pathogen. Such disruptive selection may have maintained polymorphisms of genes related to sexual fertility and stabilized gene frequencies in C. carbonum over a period of 15 years (1972-1987).