Genomic selection brings more genetic progress

What is the connection between poultry breeding and drawing blood samples from birds? Red blood cells have the hereditary (genetic) information of an animal stored in its DNA, just like it is in the cells of all other body parts. The new technique, called genomic selection, can read the genetic information and thereby learn a lot about the characteristics and quality of the animals. This information can then be used to select parents for the next generation of the breeding programme.

The goal of breeding chickens and turkeys is to mate the best animals of a generation to produce the next generation. But which ones are the best animals? The animals that perform best for the traits that the breeding company (actually its customers) finds important and likes to see most. These traits can change over time, because customers ask for an animal with different qualities. For years egg layers were selected for egg numbers, early maturity, peak production and persistency of lay and egg quality and colour. At the same time, selection attention was paid to the egg weight curve, feed conversion, body weight, hatchability and culls. All these traits were considered, more or less, strongly depending on how they fit in the breeding goal.

Emphasis on trait changes
Today robustness (especially for Europe), behaviour and feathering get more emphasis in the breeding goal while other traits get less. Not much progress can be made any more for early maturity and peak production. However persistency of lay and egg quality during the late stages of production have become much more important, according to Frans van Sambeek, director of primary breeding for ISA, the egg laying division of Hendrix Genetics. “To collect data about these traits, we have kept crossbred progeny from breeding males up to 80 weeks” he says. “Since 2009 progeny from the pure blood line is kept until 100 weeks. When you keep hens to 80 weeks then the males are older than 105 weeks. If selection is started at that late age then the selection interval is even longer. ISA also evaluates dry matter of eggs, which is a trait especially important to the egg product industry. A lower percentage dry matter indicates more water in the egg and that has to be removed to manufacture egg powder.

To establish how good an animal is, we have to measure and collect all kinds of data. Certain traits, like growth, can be measured on the animals themselves, but for many other traits, like potential egg production on males, that is not possible. In such cases the geneticist evaluates parents, half siblings and progeny. Based on all this collected data an animal is assigned a breeding value, which is an estimate of how good an animal is.

ISA uses an enormous database, because every pure blood line contains thousands of animals. In today’s poultry breeding we use statistics, calculation models and computer programmes to analyse the success of the poultry breeding company.”

Genome, chromosomes and genes
In 2004 the chicken genome was completed. This means that the chicken was genetically charted so that it was known where all the genetic information can be found. DNA (DesoxyriboNucleic Acid) is the carrier of hereditary information. DNA is passed on to the progeny. A DNA molecule consists of two long strands of nucleotide, the so called base pairs. DNA has four different bases: guanine, cytosine, adenine and thymine; these are abbreviated respectively as G, C, A and T. G and C can jointly form a base pair and so can A and T. The order of the nucleoids in a strand is called a sequence. These sequences establish the code for forming many varying proteins which perform a variety of functions in- and outside the cell. Consider for example digesting of feed as an effect on making egg shell. DNA exists in body cells in the form of chromosomes. Chickens have 39 pairs of chromosome (humans have 23) of which one pair is the sex chromosome. On each of the 39 pairs lie thousands of genes, altogether about 20.000, which contain about 1,5 billion base pairs. In other words: each gene consists of a shorter or longer DNA sequence.

Breeding out “fishy taint”
Trimethylamine (TMA) has a distinct fishy taint which can cause so called “stinky eggs” (or for cows stinky milk and for people a fishy perspiration odour). Normally the enzyme FMO3 transforms TMA into the odourless TMA-oxide which the animal passes in its urine. When FMO3 is absent TMA is not broken down which allows the taint to form in the eggs. Canola shorts is a feed ingredient that can cause fish tainted eggs because it contains sinapine which can be transformed to TMA after fermentation in the last part of the gut. FMO3 deficiency occurred only in certain lines of brown layers and then only in a small percentage of the animals. This deficiency was genetically present. Most individuals can form the enzyme FMO3, but a small number cannot.

Researchers at the University of Uppsala (Sweden) discovered that fish taint is caused in animals that have a faulty FMO3 gene, which causes the fish taint. They developed a test to prove this. The Swedish researchers formed a firm called FunboGen, which holds the patent for this test for all animal species. ISA received the exclusive rights in 2008 to perform this test for all bird species, while primary breeder Lohmann applied to perform the test for chickens.

First practical application
Both breeding companies quarrel together over these patents: Lohmann claims that Hendrix Genetics may not use the test for chickens because Lohmann has exclusive use for chickens, while Hendrix Genetics reasons that Lohmann may not apply the test on chickens because chickens are birds and they have the exclusive use to apply the test to birds. Despite this apparent stalemate both global poultry breeding companies have managed to remove the inability to destroy TMA from their brown layers. This was the first practical application of genomic selection.

The chromosomes vary from very large to very small (of course they are all microscopic). In chickens chromosome 1 is very large. This chromosome contains 15% of the hereditary traits of a chicken. Chromosome 38 contains very little. Knowing the genome of the chicken means that we know the DNA of all 39 chromosomes.

New possibilities
When the chicken genome was completed many new scientific methods were developed to apply the knowledge of this genome. It is now possible to place genetic markers on the chromosomes; kind of placing flags on the chromosomes. A very usable type of such genetic markers is the so-called single nucleoid polymorphism, or for short SNP, which is pronounced “snip”. There are 15 to 20 millions snips in a chicken. Snips cover each two sequences consisting of 100 to 200 bases. Of interest is that two joint base sequences can be different for one base only, i.e. one sequence has a C and the other sequence has a T at the same location. That may well be the reason for a difference in a certain trait. Machines were developed to read the snips, costing approximately 0.2 cents per snip.Hendrix Genetics started using this in 2005 for broilers and later for egg layers. Van Sambeek: “We started with a group of 1000 animals of which we had a lot of detail and we had a very accurate breeding value because we tested many of their progeny. Gerard Albers, director of R&D for Hendrix Genetics states: “This way we can find out what difference results from having a C or T placed at a certain position. Using the results of all 60,000 SNPs and all traits in the breeding goal of the ISA egg layers, we have created a very large table that contains the impacts of all 60,000 C-T differences on all traits. This table is then used to calculate the ‘genomic’ breeding value for current animals which only have a snip result but no trait information.”

Hendrix Genetics collaborates with several universities, one of which is Wageningen University in the Netherlands. US government organisation USDA has granted 2,5 million USD in subsidies. The advantage of this method is that a reliable estimated breeding value (EBV) of an animal on the basis of snips can be made. This can be done at a young age when the animal itself has expressed many of it’s own features and properties and of course none of its progeny.

Albers: “So we select earlier, which shortens the generation interval and thereby we can make quicker genetic progress. On the male side we progress by leaps and bounds.” Van Sambeek states: “This method is especially advantageous for features that are difficult to measure, like culling, bird behaviour, egg quality at 100 weeks of age and sensitivity to disease. For criteria like egg colour the gain is less.”

Hendrix Genetics has relegated the testing to service labs in Canada and Spain, where it has been completely automated. In the meantime many pure blood lines have been tested and the table thereof has been compiled. ISA now applies this method to several of its commercial blood lines.

The old method also remains
Although the new method of estimating breeding value (genomic selection) has been implemented, it does not mean that the old methods are thrown overboard. Albers: “We have to continue measuring on the animal itself the old way in order to conclude what the genetic potential of an animal is. Because only then can we judge with great certainty what impact a different base at a certain location makes (G, C, A or T).”Measuring production data for individual animals and their families and entering this via handheld computers into the main computer, where the estimated breeding value is calculated, will remain.”