Reliability
The reliability of genomic evaluations is an important consideration when using them to inform your breeding programme.
Reliability indicates how accurate, or the degree of confidence, the estimate for a particular genetic index is.
Typically, published reliabilities range from 20% to 99%, depending on the amount and source of information used in their calculation. The lower the reliability of the proof, the more likely it is to change as more data is incorporated into the evaluation.
Bull proof reliabilities typically improve with genomics, daughter performance, and number of herds daughters are recorded in. Female proof reliabilities typically improve with genomics, and their own performance.
Genomic reliability
Genomic testing improves the reliability of evaluations throughout the lifetime of the animal compared to traditional genetic indexes. This effect is particularly impactful for younger animals (bulls with no daughters and heifers) that otherwise have evaluations estimated from low reliability parent averages.
The reliability indicates how likely it is that the bull’s proof will change as more information becomes available. A genomic proof with a reliability of 65–70% could drop or improve when further daughter information is considered.
The figure below shows genomic proofs have a higher reliability than traditional pedigree index calculations.
The genomic proofs we produce have an identical layout to production and daughter-proven proofs, but they are flagged with ‘G’ to indicate that the proof uses DNA and parent average information to estimate the animal’s genetic potential.
Genomically tested young bulls retain this reliability gain over traditionally tested bulls, even when first-crop daughter production information is available. It is only when second-crop daughter production information is available that traditionally proven proofs close this gap in reliability.
Although genomic young bulls have a higher reliability than their traditional counterparts, you should take the same caution when using them across the herd.
We suggest using five or six genomic young bulls across your herd to ensure that if one of them changes significantly, his genetics will only contribute to a small proportion of your replacement heifers.
The figure below is a guide to the proportion of semen usage for a single bull depending on its reliability. The numbers within the bars indicate the minimum number of bulls at a given reliability recommended to generate a full bull team.
The graph shows that you should use a traditional young bull (Rel = 35%) on no more than 8% of your herd. This proportion increases to around 15% for Holstein genomic young bulls (Rel = 65%) and further for first crop daughter-proven bulls (Rel = 85%). You should use second crop daughter-proven bulls (Rel = 95%) on no more than half your herd.
Reliabilities for non-Holstein breeds will be somewhat lower because of the smaller size of the breed and informative bulls in the reference population.
The reliability is approximately doubled in genomically tested heifers, compared to a parent average proof, allowing for more confident breeding decisions to be made earlier in the animals life and thus accelerating the genetic gain of the herd.
The figure below shows an increase from 31% reliability for a parent average heifer estimate, to a 59% reliability genomic heifer proof. Reliability improvements from genomic tested are also maintained throughout a cow's lifetime.
