Restricted Maximum Likelihood Method to estimate components of genetic variation – an example in quantitative methods

Background Helminths constitute one of the most important constraints to livestock production in the tropics. Widespread infection with internal parasites in grazing animals, associated production losses, costs of anthelmintics and death of infected animals are some of the major concerns. Current control methods focus on reducing contamination of pastures through anthelmintic treatment and/or controlled grazing. In the tropics, these methods are limited by the high costs of anthelmintics, their uncertain availability and increasing frequency of drug resistance. In this situation an attractive, sustainable solution is breeding for disease resistance. Indeed, there is a large and diverse range of indigenous breeds of sheep and goats in the tropics, some of which appear to have the genetic ability to resist or tolerate helminthiasis. This was the background to a previous case study on the use of fixed effect least squares analysis to investigate the genetic resistance to helminthiasis of two indigenious breeds of sheep – Dorper and Red Maasai. As well as comparing resistance to helminthiasis between breeds it is also of interest to examine genetic variation within breeds. For this we need to use what are known as ‘restricted’ or ‘residual maximum likelihood’ (REML) procedures which are able to simultaneously estimate random and fixed effects and use the random estimates to determine heritability.