A relationship matrix including full pedigree and genomic information

Area of Science:

• Animal Genetics
• Quantitative Genetics
• Bioinformatics

Background:

• Dense molecular markers are increasingly used in genetic evaluations, often requiring a two-step process.
• Current methods using pseudo-data lead to bias and information loss.
• Directly substituting pedigree relationships with genomic relationships causes inconsistencies.

Purpose of the Study:

• To develop a method for directly incorporating genomic information into genetic evaluations.
• To address biases and information loss associated with two-step procedures.
• To create a unified relationship matrix that combines pedigree and genomic data.

Main Methods:

• Proposed a pedigree-genomic relationship matrix (H) by conditioning genetic values of ungenotyped animals on genotyped animals.
• Utilized a selection index approach to link genotyped and ungenotyped individuals.
• Ensured the matrix H is (semi)positive definite by construction.

Main Results:

• The proposed matrix H integrates genomic information across all individuals, including ungenotyped ones.
• The new approach avoids the incoherencies and biases of naive methods.
• Matrix H is suitable for efficient computational algorithms like preconditioned conjugated gradients.

Conclusions:

• The pedigree-genomic relationship matrix (H) offers a more accurate and consistent approach to genomic evaluation.
• This method effectively transmits genomic information throughout the population structure.
• The developed matrix is computationally viable for large-scale genetic analyses.