Allele coding in genomic evaluation.

Area of Science:

• Animal genetics
• Quantitative genetics
• Statistical genomics

Background:

• Genomic data aids genetic evaluation in animal breeding.
• Two main approaches exist: estimating marker effects or direct estimation using genomic relationship matrices.
• Allele coding assigns values to regression coefficients, with common methods including 0,1,2 coding and centered coding.

Purpose of the Study:

• To investigate the effects of different allele coding methods on genetic evaluation inference.
• To compare marker-based and equivalent models under various allele coding schemes.
• To assess the impact of allele coding on the reliability of genomic breeding values and computational efficiency.

Main Methods:

• Utilized both marker-based and equivalent genomic models.
• Employed restricted maximum likelihood and Bayesian inference methods.
• Analyzed theoretical derivations and simulation studies to evaluate parameter estimates and model performance.

Main Results:

• Parameter estimates and marker effects are consistent across allele coding methods in marker-based models with a fixed general mean.
• Genomic breeding values are independent of allele coding when the general mean is incorporated.
• Reliabilities of genomic breeding values vary with allele coding, and centered coding improves Markov chain Monte Carlo algorithm mixing.

Conclusions:

• Allele coding methods yield consistent inference in both marker-based and equivalent models if a fixed general mean is included.
• The choice of allele coding impacts the reliability of estimated genomic breeding values.
• Centered allele coding provides numerical advantages for Bayesian inference using Markov chain Monte Carlo methods.