Estimation of genetic parameters: it takes three to tango

Abstract

Estimation of genetic parameters is one of the classic tasks of quantitative genetics. For continuous traits, this requires partitioning of observed variation into its genetic and environmental components. I clearly recall my sense of wonder when first introduced to the concept: Given records on relatives and an estimation of the covariance between them, we can infer how heritable our traits of interest are. Estimation has three pillars: First - and foremost - the data, which comprise not only phenotypic records, but also relationship (pedigree) information. Secondly, inference requires a model of analysis, summarizing our assumptions, including those on the mode of inheritance. Thirdly, we need a statistical framework and appropriate strategies to carry out the computations involved.