Measure of Multilocus Correlation as a New Parameter for Study of Plant Pathogen Populations

Abstract

ABSTRACT The measure of multilocus correlation for diallelic loci was developed to augment use of other diversity indices in the study of multilocus structure of populations. This measure provides information not revealed by other parameters for measuring multilocus association and linkage disequilibrium among pairs of loci. Relationships between the measures of multilocus correlation and association and all commonly used indices for diversity within populations are also described. A number of hypothetical examples demonstrate that the measure of multilocus correlation describes unique aspects of multilocus structure of populations compared with the measure of multilocus association and phi coefficient of association. The measure of multilocus correlation can be used for analysis of independence of differentiating characters for a given population-the necessary condition of valid applications of the bootstrap method across differentials. Comparisons of the measures of multilocus correlation and association with the measures of diversity within population show that they reveal different aspects of population structure. As an adjunct to standard diversity indices, the measure of multilocus correlation may reveal subtle differences in diversity within populations even if the standard index fails to distinguish between the populations. A new measure of population uniformity also was developed to characterize another aspect of diversity within populations not detected by standard indices.