Development and Commercialization of Software for Genetic Improvement Programmes: a Case Study

The development of computers within the last 20 years has been dramatic. Whilst computer scientists may have been making predictions of such changes, it would have been hard for livestock breeders to appreciate fully that, within a short space of time, populations of millions would have individual breeding values calculated using complex and demanding statistical techniques. Examination of pre-1990 research indicates the extent of computational development since. Many techniques that are common and accepted in research today were rare, e.g. simple stochastic simulation of populations and Bayesian analysis of complex data using Monte-Carlo Markov Chains. Even providing a standard error for a variance component was a challenge! This surge in computing power has prompted the development of concepts to remove important areas of ignorance in our understanding of genetic improvement schemes, and it is not surprising that this same surge has had a profound impact on their operational management.