Estimation and Tests of Mendelian Inheritance in Small Samples of Human Families

Estimation of parameters in human genetics is usually performed by Maximum Likelihood methods for large samples relying on asymptotic results. With the aim of evaluating the importance of the bias resulting from sample size, two sets of human family data have been analysed. The bias is not very important, but programs have been elaborated, giving better estimations. The tests of hypotheses from these estimations have also been considered.