Genetic selection stopped being a matter of picking the best-looking bird a long time ago. What changed most recently is not the ambition, it is the speed: a breeding programme can now turn a generation in seven months instead of thirteen, which means every choice of objective compounds twice as fast.
Key takeaways
- Genomic selection entered commercial layer breeding in 2013, with SNP panels growing from 6 000 markers in 2007 to 600 000 by 2013 (Lohmann Breeders).
- Reported gains in prediction accuracy against pedigree-based estimates run from 20 % to 70 % depending on the trait, with the largest benefit on reproduction and livability.
- In a proof-of-concept programme, the genomic line produced a new generation every 7 months against 13 for the conventional line.
- Selecting hard on growth has documented costs. Comparative work backed by the RSPCA reports higher mortality and more breast muscle myopathies in conventional fast-growing strains than in a slower-growing commercial breed.
What genomic selection actually changed
Genomic selection does not replace measurement, it moves it. Instead of waiting for a bird or its progeny to express a trait, breeders genotype the candidate and predict its breeding value from markers calibrated on a reference population. Hy-Line International implemented it in layer breeding in 2013, and the numbers behind that shift are documented: SNP chips moved from 6 000 markers in 2007 to 12 000 in 2011, then 42 000 and 600 000 in 2013, while the number of genotyped birds went from a few hundred per line to tens of thousands.
Two effects follow. The first is accuracy, reported at 20 % to 70 % better than pedigree-based estimates for reproductive traits and livability in broiler breeding. The second is time. Because a candidate no longer has to wait for its own record, the generation interval collapses, and in the proof-of-concept cited by Lohmann Breeders the genomic line ran four generations in the three years the conventional line needed for two.
The traits pedigree selection could never reach
The traits that gain most are precisely those a classical programme handles badly: sex-limited traits such as egg production, which a male never expresses; traits recorded late in life, such as persistency of lay; traits with low heritability, where individual performance is a poor signal; and traits that cannot be measured on the selection candidate itself without harming or destroying it.
Disease resistance sits in that last group and deserves a caveat. Selecting for a general robustness phenotype, mainly through livability and mortality records, is established practice. Selecting for resistance to a specific pathogen is a much narrower claim, and one that should always be read with the challenge model, the strain and the environment attached. Genetics reduces susceptibility at population level; it does not exempt a house from biosecurity or from a vaccination programme designed with your veterinarian.
The trade-off the sector can no longer set aside
Decades of selection weighted towards growth rate and breast yield changed broiler morphology, and a 2024 analysis published in Frontiers in Animal Science links that shift to consequences for musculoskeletal, cardiovascular and behavioural integrity. The same analysis notes that adding health and welfare traits into fast-growing breeding programmes has so far had only a limited corrective effect, which is the uncomfortable part of the finding.
The comparative field data point in the same direction. Work backed by the RSPCA reports that Cobb 500, Ross 308 and Hubbard Flex birds showed higher mortality, more white striping and wooden breast, and poorer leg, joint and plumage health than a slower-growing commercially viable breed. In another commercial comparison, slower-growing Hubbard Redbro birds were slaughtered 5.5 days later at a lower average weight, with lower mortality, fewer culls and fewer carcasses downgraded at the abattoir.
| Criterion | Conventional fast-growing lines | Slower-growing lines |
|---|---|---|
| Time to slaughter weight | Shortest cycle, more crops per house per year | Around 5.5 days longer in the Redbro comparison, at lower weight |
| Mortality and culls | Higher in the published comparisons | Lower mortality, fewer culls |
| Carcass quality | More breast myopathies, more downgrades | Fewer downgrades at the abattoir |
| Feed and land per kg of meat | Best conversion, lowest cost per kg | Higher feed use and longer occupancy per kg produced |
| Market access | Standard indoor specifications | Required by EU organic rules, and by welfare commitment schemes |
Where the rules sit in 2026
Two frameworks bear on a breed decision in Europe, and they are of different natures. Council Directive 2007/43/EC is law: it caps stocking density at 33 kg/m² in general, allows a maximum of 39 kg/m² where additional environmental requirements are met, and permits up to 42 kg/m² where cumulative daily mortality stays low. It regulates the house, not the genetics.
The European Chicken Commitment is a private, NGO-backed standard rather than legislation, and it does address genetics. Signatories undertook to move to its specifications by 1 January 2026, with a maximum density of 30 kg/m² and the use of named breeds such as Hubbard Norfolk Black, JA757, 787, 957 and 987, Rambler Ranger, Ranger Classic and Ranger Gold, or of any breed meeting the RSPCA Broiler Breed Welfare Assessment Protocol. Reporting published in 2025 by World Animal Protection found that a large share of signatories were not on track at that deadline, so a buyer specification citing the commitment should be checked against actual delivery rather than assumed.
How to read a genetics claim before you sign for it
This is where a small habit saves money. Google Scholar is a search index, not a source. A supplier citing “studies on Google Scholar” or “the literature” has told you nothing until you can reach the underlying paper, and the same applies to a bare mention of a DOI without the reference attached to it.
When a line is presented with a performance or resilience claim, the questions that matter are short:
- Which trial population and which flock size? A pen trial and a 30 000-bird house are not the same evidence.
- Which housing system and climate? Resilience measured in a temperate closed house transfers poorly to an open-sided house in a hot season.
- What was the comparator, and was it a current commercial line or an outdated one?
- Which traits moved, and which traits were traded to move them? A gain in growth alongside a silence on leg health is an answer in itself.
- Is the figure published in the management guide for that specific line, or does it come from marketing material?
Straight answers on genetics
Does genomic selection mean genetic modification?
No. Genomic selection reads existing natural variation in the genome to predict which birds to breed from. Nothing is inserted or edited. The birds produced are the result of conventional mating, chosen with better information.
Can a farm select its own lines?
For a commercial production flock, no, and the attempt usually costs more than it returns. Reciprocal crossbreeding programmes run at pyramid level over many generations and large populations. Conservation of local and heritage breeds is a legitimate and different objective, generally supported by breed associations and public programmes.
Is genetic diversity a real concern?
It is, and it is one of the ethical issues acknowledged within the sector itself. Commercial poultry production depends on a small number of lines held by a handful of breeding companies, which makes conserved populations and gene banks a form of insurance rather than a heritage exercise. Whether current safeguards are sufficient is genuinely debated, and we will not pretend to settle it here.
Which breed should a producer choose?
The one that matches the outlet, the housing and the specification you are contracted to, in that order. Selection at farm level is a purchasing decision, not a breeding one, and it is best made against the management guide of each candidate line. Our comparison of laying hen breeds by production profile sets out how we approach that arbitration.
From the breeding index to your order form
Genetics only becomes an operational decision at the moment you choose a line for a specific outlet: eggs, meat, or both.
Sources: Lohmann Breeders, Genomic selection in layer and broiler breeding; Frontiers in Animal Science, analysis of the welfare of fast-growing and slower-growing strains of broiler chicken (2024); RSPCA-backed comparative studies of conventional and slower-growing broiler breeds; Council Directive 2007/43/EC, Article 3 and Annex II; European Chicken Commitment published policy; World Animal Protection, The Pecking Order (2025). Consulted August 2026.
Published previously, fully revised on 16 August 2026. General operational guidance for poultry professionals. It does not replace the assessment of the veterinarian responsible for your flock or the management guide published for your genetics.