Transformation of multitrait to unitrait mixed model analysis of data with multiple random effects

Area of Science:

• Statistics
• Quantitative Genetics
• Animal Breeding

Background:

• Multitrait mixed model analyses are computationally intensive.
• Existing methods require substantial computational resources (CPU time and memory).
• Simultaneous diagonalization of covariance matrices is a key assumption in some models.

Purpose of the Study:

• To present an algorithm for transforming multitrait analysis into a unitrait analysis.
• To simplify the setup and solution of mixed model equations.
• To reduce computational burden in multitrait analyses.

Main Methods:

• Developed an algorithm to transform multitrait mixed models into unitrait models.
• Restricted models to those with simultaneously diagonalizable covariance matrices for random effects.
• Applied a transformation based on the common principal component concept.

Main Results:

• Transformed multitrait analysis simplifies to solving unitrait equations for transformed traits.
• Significantly reduces programming complexity for mixed model analysis.
• Drastically decreases central processing unit time and computer space requirements.

Conclusions:

• The proposed algorithm offers a computationally efficient alternative for multitrait mixed model analysis.
• This method simplifies complex statistical analyses, making them more accessible.
• Reduces computational costs, enabling more extensive analyses in quantitative genetics and breeding.