New Bayesian Methods for Genomic Selection
Author Information
Author(s): David Habier, Rohan L. Fernando, Kadir Kizilkaya, Dorian J. Garrick
Primary Institution: Iowa State University
Hypothesis
Can new Bayesian methods improve genomic prediction by treating the prior probability of SNP effects as unknown?
Conclusion
BayesCπ is a promising method for genomic prediction that accounts for computing effort and uncertainty in the number of QTL.
Supporting Evidence
- BayesCπ showed better accuracy in distinguishing QTL from SNPs as training data size increased.
- The accuracy of genomic estimated breeding values (GEBVs) improved with larger training data sizes.
- BayesA was found to be a good model choice for genomic prediction at the current SNP density.
Takeaway
Scientists created new methods to better predict traits in animals using genetic data, which helps in breeding better livestock.
Methodology
The study used simulations and real data from North American Holstein bulls to compare the accuracy of different Bayesian methods for genomic prediction.
Limitations
The methods may overestimate the number of QTL and are sensitive to training data size and SNP density.
Participant Demographics
The study involved 7,094 North American Holstein bulls.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website