Comparison of individual assignment methods and factors affecting assignment success in cattle breeds using microsatellites

Abstract

Comparisons of seven different individual assignment methods based on likelihood and distances were carried out with four different sets of data, which varied both in number of populations and genetic divergence. Based on 27 microsatellites genotyped in eight cattle breeds (Icelandic and seven Norwegian), 28 hybrid populations were simulated. Factors affecting individual assignment success, such as number and divergence of populations, sample size and number of loci, were assessed from actual and simulated data. The Bayesian, frequency and Nei-minimum methods performed more or less similarly. Individual assignment success depended mainly on population number and divergence (FST). Higher success was observed at high level of divergence among populations and at low number of source populations considered. With eight pure breeds and 27 loci considered the assignment success rate ranged from 55 to 70%. Generally, assignment success increased with increasing number of loci and/or sample size.