Lars Rönnegård: Applications of hierarchical likelihood methods in genetics

Sammanfattning: Finding how robust individuals are to environmental variations has been an important question in genetics over the last century. In the presentation I will show how modelling of the residual variance in linear and linear mixed models can be used to study phenotypic robustness in genetics. Going from a relatively simple model to more complex ones, the concept of hierarchical likelihood methods will be introduced. Lee and Nelder (1996) defined the hierarchical likelihood, which can be used for both estimation and inference of generalized linear models with random effects. It combines modelling flexibility with computational speed and thereby allows analysis of large data sets. Results from analysis of large data sets in genetics will be presented.