When duck broiler selection started decades ago, the focus was on the growth traits, especially body weight and daily gain. Applying selection on those traits gave the opportunity to meet the expectations of the farmers and make the duck business more efficient. Nevertheless, breeding companies do not focus only on the final traits but on all traits that impact the duck broiler value chain. Which means also reproductive traits: egg production, fertility, hatchability… Genetic selection brings a long-lasting solution to the duck value chain but requires key points, such as the right traits definition, accurate measurements, and an appropriate statistical model.
Data collection
In the Orvia breeding programme, different traits will be individually recorded in the genetic database. This is an in-house database that has been specially developed for pure line breeding companies’ purposes. Next to this database, devices have been developed, adapted to fulfil data collection (handheld, barcode, RFID, ultrasonic device, image analyser use). First of all, egg production is individually registered. All females have a known egg production. When Orvia talk about egg production, they are talking about hatching eggs. Indeed, cracked eggs, small eggs, and double yolk eggs are not counted because they will never give ducklings. Every year, they talk about millions of eggs that are recorded. Egg weight and shell strength are also monitored. Different egg weights per female are measured to ensure that the egg weight stays within a good range. Shell strength measurement will allow minimising the number of broken eggs, which means maximising the number of hatching eggs.
Females are not the only ones to be monitored on reproductive traits. The semen quality of males is scored several times during their lives. For such tests, microscopes and spectrophotometers can be used. The quality and quantity of semen are recorded (volume, motility, percentage alive, concentration). Fertility and hatchability are also monitored. The fertility data are recorded during candling (Fig. 1). They do not limit themselves to this ‘fertility’. True fertility is also measured. A breakout session is done; all clear eggs are broken to determine whether there has been an embryonic development and, if so, when embryonic development stopped. For each hatch, the hatchability results are registered in the database per dam. Only ducklings of good quality – perfectly hatched without phenotypic defects – are taken into account for this purpose.
Orvia also perform a semen storage duration test. After the last artificial insemination, they recorded the number of fertile eggs, which mixed both traits: egg production and semen storage duration.
Some traits may be difficult to measure, for example, sexual behaviour. Despite video recordings made during experiments, this is a complex trait to be considered. But their GP flocks allow them to naturally favour correct sexual behaviour. Indeed, males with the best semen quality and the best sexual behaviour will produce the majority of the PS. Next to all the reproductive traits, growth traits such as body weight, daily gain, feed intake, feed efficiency, residual feed intake, feeding behaviour, walking score, breast thickness, conformation score, and other slaughtering traits are also monitored. All measurements are done following a strict procedure to be able to collect accurate data for all animals on a one-to-one basis. The broiler duckling is the final product of four-way pure line crosses. All measurements described above are applied to all pure lines.
Heritability, correlation, and genetic potential estimation
For all traits mentioned above, genetic parameters and genetic breeding values are estimated. The goal of these computation steps is to get a more accurate genetic potential for each animal. One of the genetic parameters is heritability. It represents the part of the phenotypic variability explained by genetic variability. Its value fluctuates from 0 to 1. For reproductive traits, heritability is from 0.14 to 0.55, depending on the traits and pure lines. Unfortunately, growth traits are sometimes negatively correlated with reproduction traits. Correlations are from -0.66 to 0.95. A quick analysis could conclude that improving reproduction traits will decrease the performance of the ones which are negatively correlated. This is false, thanks to the selection process. Part of the reproductive traits are so-called late traits. This means that a part of the data is collected after pure line renewal (example: late egg production). And it reduces accuracy.
To tackle this problem, one option is to renew the pure line just at the end of the laying cycle. But this will increase the generation interval and reduce the genetic response. To avoid this situation, Orvia is using genomic selection. Indeed, genomic selection gives the opportunity to have a good accuracy of breeding values also for traits that are only measurable at older ages. Since 2020, thousands of birds have been genotyped on a 60K SNP-chip in their routine breeding programme. The SNPs (DNA markers) used have been chosen especially for their lines to be sure that all SNPs will be informative. But do not think that genomics will substitute the phenotypic data collection. The link between phenotype and genetic markers can easily change by DNA reorganisation. That is why Orvia measures, is measuring and will measure all its pure lines for all the different traits.
Selection process
The Orvia breeding program is based on a balanced selection process. Their goal is not to focus on reproduction traits during a few generations with huge genetic improvement and stabilising/losing in growth traits, then doing the opposite during the next generations. Orvia prefer to improve both traits regularly at the same time. The negative correlation mentioned above is the average of the population. Within a population, there is diversity of animals. In Fig. 2, the genetic potential of trait 1 is on the X-axis, and the genetic potential of trait 2 is on the Y-axis. Each dot is an animal. The negative correlation is clearly visible. Nevertheless, if they observe animals that are at the right of the red line and above the blue one, they have an average genetic potential of trait 1 higher than the population and the same for the genetic potential of trait 2.
By selecting the population in this area generation after generation, they meet the expectations of the duck broiler value chain. This will be possible only with a large population size and high selection intensity, as Orvia has. This also implies that inbreeding is monitored carefully.
P=G+E
Last but not least, what duck producers get is what is called phenotype (a measurable value of a trait). This phenotype is the result of the genetic potential and environmental effects. With accurate data collection, a large population size, and a good generation interval, the genetic potential is improved by breeding companies as described in this article. Environmental effects are managed by duck producers, like density, temperature, humidity, air flow, litter quality, and feed quality. This is why following management guidelines in relation to local specificity is key. The best performances at all stages will be achieved when good genetic potential meets the right management of the flock. Today, with their balanced breeding programme, their PS can produce 252 ducklings in 52 weeks of lay with a final product potential of 3.74 kg and 1.8 FCR at 42 days in normal continental weather conditions.
