Six procedures for measuring the admixture proportions in a hybrid population from gene frequency information are evaluated empirically against the completely ascertained data generated by computer simulation. Analysis of the standard errors of the estimates indicates that the maximum likelihood solution of Krieger et al. offers the most reliable estimates when sampling error in the ancestral frequencies is not great.
Korey, Kenneth A.
"A Critical Appraisal of Methods for Measuring Admixture,"
3, Article 11.
Available at: https://digitalcommons.wayne.edu/humbiol/vol50/iss3/11