SIMULATION OF MARKER ASSISTED SELECTION IN HYBRID POPULATIONS

A computer model is developed that simulates Marker Assisted Selection (MAS) in a population produced by a cross between two inbred lines. S election is based on an index that incorporates both phenotypic and mo lecular information. Molecular markers contributing to the index and t heir relative weights are determined by multiple regression of individ ual phenotype on the markers. The model is applied to investigate the efficiency of MAS as affected by several factors including total numbe r of markers in the genome, number of markers contributing to the inde x, population size and heritability of the character. It is demonstrat ed that selection based on genetic markers can effectively utilize the linkage disequilibrium between genetic markers and QTLs created by cr ossing inbred lines. Selection is more efficient if markers contributi ng to the index are re-evaluated each generation than if they are eval uated only once. Increasing the total number of markers in the genome as well as the number of markers contributing to the index does not ne cessarily result in a higher efficiency of selection. Moreover, too ma ny markers may result in a weaker response to selection. Population si ze is shown to be the most important factor affecting the efficiency o f MAS.