Comparison of four direct algorithms for computing inbreeding coefficients

Abstract

ABSTRACTOne widely used measure in genetic analyses of livestock species is the inbreeding coefficient. Computation costs of inbreeding coefficients are of importance for large populations. Very recently, a ‘direct’ method for computing inbreeding coefficients was modified using a relatively efficient method to construct lists of ancestors in which the integrity of chronological order within each ancestral path is kept. Using simulated data, the computational efficiency of the recently modified algorithm was investigated by comparing it with three other algorithms – the original direct algorithm, its previously modified algorithm and another direct algorithm. The recently modified algorithm became considerably faster than the original algorithm and its previous modification when the number of generations increased, and it was fast relative to the other direct algorithm when the average number of generations and family size were not large. When only animals born in the most recent year were evaluated with the knowledge of inbreeding coefficients for their ancestors, the recently modified algorithm outperformed the other algorithms, indicating its possible advantage in updating situations.