Efficient computation of genotype probabilities for loci with many alleles: II. Iterative method for large, complex pedigrees

An algorithm for computing genotype probabilities for marker loci with many alleles in large, complex pedigrees with missing marker data is presented. The algorithm can also be used to calculate grandparental origin probabili ties, which summarize the segregation pattern and are useful for mapping qu antitative trait loci. The algorithm is iterative and is based on peeling o n alleles instead of the traditional peeling on genotypes. This makes the a lgorithm more computationally efficient for loci with many alleles. The alg orithm is approximate in pedigrees that contain loops, including loops gene rated by full sibs. The algorithm has no restrictions on pedigree structure or missing marker phenotypes, although together those factors affect the d egree of approximation. In livestock pedigrees with dense marker data, the degree of approximation may be minimal. The algorithm can be used with an i ncomplete penetrance model for marker loci. Thus, it takes into account the possibility of marker scoring errors and helps to identify them. The algor ithm provides a computationally feasible method to analyze genetic marker d ata in large, complex livestock pedigrees.