Finite sample properties of maximum likelihood estimates of the recombination fraction in double backcross matings in man.

The properties of the maximum likelihood estimator (MLE) of the recombination fraction, theta, based on various numbers and sizes of sibships derived from double backcross matings are exactly explored where the coupling phases are presumed equally likely (which for two-generation data is generally the case). The results indicate that for large values of theta the expectations are severely biased. The bias, variance, and measures of normality of the MLE behave erratically for small sizes and numbers of sibships. The implications for chromosome mapping are discussed.