Maximum likelihood estimation by counting methods under polygenic and mixed models in human pedigrees

Abstract

For pedigree data, the maximum likelihood estimates of the parameters in polygenic and mixed models are derived analytically although not in closed form but in terms of "counting equations" allowing an iterative solution. Likelihood computations, tests of significance, and tests of goodness of fit are presented. Accelerating the (linear) rate of convergence by a very simple method is demonstrated.