A note on the effect of observations with missing data on genetic correlation estimates.

Area of Science:

• Animal Breeding and Genetics
• Quantitative Genetics

Background:

• Covariance components are estimated using variance components of sums and individual variables.
• Missing data can introduce bias in variance and covariance component estimation.
• Previous research suggests eliminating missing data to mitigate selection bias.

Purpose of the Study:

• To evaluate the impact of missing data on covariance component and genetic correlation estimates.
• To compare estimation methods with and without the exclusion of missing data.

Main Methods:

• Randomly deleted varying proportions of records for one variable in paired analyses.
• Analyzed both sample beef cattle data and simulated datasets.
• Compared genetic correlation estimates obtained with and without excluding missing data.

Main Results:

• Excluding observations with missing data resulted in more accurate genetic correlation estimates.
• Including missing data led to common estimates of genetic correlation outside the valid parameter space.
• The accuracy of covariance component estimates was also improved by data imputation.

Conclusions:

• The method for estimating covariance components should only be applied after eliminating observations with missing data.
• Excluding missing data is crucial for reliable genetic correlation and covariance component estimation in animal breeding.
• This approach enhances the precision of genetic parameter estimation in populations with incomplete records.