A MIXED MODEL FOR ANALYSES OF DATA ON MULTIPLE GENETIC-MARKERS

Data on a genetic marker linked to a gene affecting an important trait could help us to estimate breeding values for that triat more accurately. The accuracy is enhanced if many genetic markers are used and if important genes are bracketed by two markers. A mixed model for analysis of this type of data is presented. The model is appropriate for an arbitrary pedigree structure in an outbreeding species. It uses a "relationship" matrix among marked chromosome segments or QTL alleles. By using an analysis analogous to a reduced animal model, the number of effects to be estimated can be greatly reduced. A grouping strategy that can account for crossbreeding and linkage disequilibrium between markers and QTL alleles is included in the model. For analyses of a cross between inbred lines the model can be simplified. This simplification shows clearly the relationship of the mixed model analyses to multiple regression models used previously. The simplified model may also be useful for some experiments in outbreeding populations.