A new catalyst in plant breeding
Today breeders select material for a new variety based on different trial data that has been collected during the growth season. Such data may cover dry matter yield, seed yield, heading date, and disease resistance. Typically, only the top 5-10% are selected for further breeding while the rest is discarded. Same selection criteria are applied second year on the next generation and so forth until the improvements that will qualify the variety for National listing has been obtained. This system is now facing a technological development, which may lift the trait values to unseen heights.
The technology, called "Genome Wide Selection" (GWS), bases itself on the genetic potential, which is hidden in the plant genome. Shortly, GWS consists in selecting breeding material on the basis of the plants DNA code instead of their trial data. In order to do so first task is to determine, which of the DNA variations contribute positively or negatively to various crop traits. This is done by aligning DNA codes from many plants with their trial data from the field. For more than a decade DLF-TRIFOLIUM has recorded both trial data and seeds from thousands of breeding families, which are well suited for such an alignment. Once an association is established, the genetic potential for any trait can be calculated for each plant, and those with the highest breeding value can be selected. This adds completely new perspectives and will alleviate some of the limitations associated with the current breeding system.
One of the current limitations lies in the ability to select for more traits simultaneously. Today the development of a new variety with superior yield requires at least 500 field plots. If the new variety is also going to be superior in disease resistance the breeding process requires 500 x 500 (= 25,000) field plots. This is an impossible task, which in practice means that you can select for one trait (typically yield) and then only hope to find variation for other traits in your selected material. GWS can overcome these limitations because it can dissect, which part of the genome controls different traits. Thereby it is possible to select plants with the highest breeding values for both traits. In the long term GWS may also reduce the number of expensive field trials because selections are made on calculated breeding values derived from genomic data. Value precision will increase with increasing number of breeding lines in the underlying calculation model.
This technology has been implemented in cattle breeding with great success and the best breeding bulls are now selected even before one year’s age based on genetic breeding value. Prior to GWS, data on these animals and their offspring was collected for 5-6 years before it could be concluded, which bull had the greatest potential.
The aim of this project is to develop and tailor GWS for grass breeding. In contrast to cattle breeding grasses are bred in families (consisting of several plants) and not individuals. This difference raises several scientific and technical challenges, which the partners will seek to solve. One challenge is to develop a new genotyping method for determination of family DNA profiles. Another is to tailor new statistical models to outbreeding organisms based on quantitative genetics. It is expected that developments made in this project will pave the way for new innovative solutions in other areas of genomic prediction.
Genome and GWS
The genome in most organisms consists of long DNA chains (Chromosomes), which harbors genes that control different traits. The building blocks in DNA consist of four different ribonucleic acids, which are designated by A, T, G, and C. Ryegrass has seven chromosomes and contain approximately 2.7 billion bases and 40,000 genes.
It is the differences in the DNA codes that accounts for the observable differences between to genetically different individuals. Some genes are controlled by a single gene but most are controlled by multiple genes.
In order to apply GWS in breeding it is not necessary to know all DNA differences between the families but preferably there should be at least 100,000 or more.