A brief history of statistical developments in animal breeding 1 - from Galton to Bayes

Published on January 31, 2017   29 min

Other Talks in the Series: Statistical Genetics

Please wait while the transcript is being prepared...
0:00
Hello, my name is Daniel Gianola. I'm a Professor at the University of Wisconsin-Madison, and I'm also a Hans Fischer Senior Fellow at the Technical University of Munich, in Germany. And today, I will be talking about "A Brief History of some Statistical Developments in Animal Breeding". I will cover about 100 years of developments.
0:24
So the first question is, "What is animal breeding?" By this, we mean the genetic improvement of farm animals mostly, animals that serve and enrich mankind. The question is, "What individual is best?" For that, we need to define some "merit function", which typically will depend on several traits. The same happens in plant breeding. For example, some variety of wheat will depend not only on grain yield but also on dry matter content, resistance to disease, resistance to wind, and some other attributes. In order to decide what animals or individuals are best, we take measurements, such as milk production in dairy cows, growth rates in animals. We have genealogies or pedigrees that trace animals according to lines of descent. More recently, molecular markers that we know as SNPs, S-N-Ps, single nucleotide polymorphisms. In the future, full DNA sequences, and also post-genomic data, such as gene expression, level of methylation, which are now available for experimental purposes; but have not been used for predicting breeding value of animals, yet. Now, the data that we work with in animal breeding are largely observational and retrospective. They consist of farm records that have been collected directly on the farm. And there is a notorious absence of randomized experiments, as in human medicine, mostly because it's impossible to carry out large animal experiments, as it is impossible or very difficult to randomize patients into clinical trials. Then, the question of evaluating the merit of animals becomes largely a statistical one. And then, once we have evaluation of the genetic merit of the candidates. We decide, which will be selected? How we are going to be mating the selected animals? Are we are going to be mating relatives with relatives? Or we are going to avoid mating relatives?
Hide

A brief history of statistical developments in animal breeding 1 - from Galton to Bayes

Embed in course/own notes