Genetic evaluation models turn performance records, pedigrees and now DNA information into estimated breeding values (EBVs) that rank animals as parents. The two main approaches compared in swine are BLUP, which uses phenotypes and pedigree, and single-step genomic evaluation, which blends genomic and pedigree data in one analysis. This article explains each model in plain language, what they require, how they differ in accuracy and why most breeding companies are moving toward single-step methods. It also gives practical advice on how to read and use EBVs.
What an EBV tells you
An estimated breeding value is an estimate of the genetic merit an animal passes to its offspring, expressed in the unit of the trait or as an index. It separates genetic effects from environment, such as pen, season or herd. EBVs come with an accuracy figure showing how reliable the estimate is. A high EBV with low accuracy may change as more data arrives. Use EBVs to compare animals within the same evaluation, rather than between different companies or systems.
BLUP: the standard
BLUP stands for best linear unbiased prediction. It uses a statistical model that includes records on each animal, records on relatives through the pedigree, and environmental factors such as contemporary group, sex and age. Because it uses the relationship matrix, information from relatives, sires and dams improves accuracy. BLUP has driven swine genetic progress for decades and is the foundation of most national and company evaluations. Its main limit is that pedigree relationships are averages and cannot capture which chromosome segments an animal actually inherited.
Genomic data and the step approach
DNA markers (SNPs) allow the actual genetic relationship between animals to be measured. Early genomic evaluation used multiple steps: BLUP was run first, results were turned into pseudo-phenotypes for genotyped animals, marker effects were estimated and then combined with pedigree EBVs. This approach loses information, requires complex weighting and can bias results. See SNP Chips for Swine: Density and Application and Genomic Selection in Swine: Current Adoption and ROI.
Single-step evaluation
Single-step genomic BLUP (ssGBLUP) combines the pedigree relationship matrix and the genomic relationship matrix into one blended matrix. Genotyped and non-genotyped animals are evaluated together in one analysis, using all phenotypes, pedigrees and genotypes. This avoids the need for pseudo-phenotypes and reduces bias, especially when selected animals are genotyped. It handles many traits and a large number of animals but requires strong computing and careful data management.
Comparison
| Feature | BLUP | Single-step |
|---|---|---|
| Data used | Phenotype and pedigree | Phenotype, pedigree and genotype |
| Accuracy for young animals | Moderate | Higher, if genotyped |
| Complexity | Lower | Higher |
| Genotypes needed | No | Yes, for some animals |
| Bias from selection | Handled by model | Handled, with care in blending |
| Typical use | Smaller programs | Major breeding companies |
Why accuracy rises
Genotypes allow a piglet to be evaluated at birth rather than after her own or her offspring's records exist. For traits that are expensive, late in life or hard to measure, such as sow longevity, litter traits, feed efficiency or disease resilience, genomic information adds a lot. The gain is largest when the reference population is large and closely related to the candidates. Single-step also corrects mistakes in pedigree, because genotypes reveal true relationships, see Parentage Verification: DNA Testing in Multi-Sire Groups.
Practical implications for producers
- Ask your supplier which model they use and how often evaluations are updated.
- Treat genomic EBVs as more accurate for young animals but still estimates.
- Compare animals only within the same evaluation and breed line.
- Use indexes that combine traits in line with your economic goals, see Maternal Line Selection: Balancing Litter Size and Robustness.
- Do not ignore structure, health and temperament when reading EBVs.
- Keep your own records, because they may feed into the evaluation.
Limits and pitfalls
Evaluations are only as good as the data. Poor recording, wrong pedigrees, small contemporary groups or biased genotyping lead to errors. Genetic by environment interaction means that a line evaluated in a nucleus herd may perform differently on your farm, see Nucleus vs. Multiplication Herds: Genetic Lag. Rapid selection can reduce genetic variation and increase inbreeding, see Inbreeding Depression in Closed Herds. Models are updated as new methods become available, so check with your supplier.
Data you can contribute
Better data make better evaluations, and producers who supply records are part of the process. Accurate sire and dam identification, farrowing dates, litter sizes, number weaned, culling reasons and leg or teat scores all add value. Consistent contemporary groups, meaning animals raised together in the same barn and period, are essential, so keep pen and batch information. If you participate in a genotyping program, handle tissue samples carefully and label them correctly, since mix-ups create errors that are difficult to find. Ask for feedback reports, because they show how your herd compares and where selection should focus. Over time, herds that record well tend to get more useful EBVs and more accurate decisions.
FAQ
What is the difference between BLUP and single-step?
BLUP uses performance and pedigree data, while single-step also uses DNA genotypes in a single combined analysis.
Do I need genotyped animals to use EBVs?
No, you can use ordinary EBVs, but genotyping improves accuracy for young animals.
Can I compare EBVs from different companies?
Generally no, because each evaluation has its own base and model.