Analysis of a genetically structured variance heterogeneity model using the Box-Cox transformation

Area of Science:

• Quantitative genetics
• Statistical genetics
• Animal breeding

Background:

• Genetic factors influencing environmental variance are increasingly recognized.
• Skewness in data distributions can lead to misleading inferences in variance component models.
• Parametric models often assume normality, which may not hold for biological data.

Purpose of the Study:

• To investigate the effect of data scale on inferences from genetically structured heterogeneous variance models.
• To extend these models using Box-Cox transformations to account for data asymmetry.
• To re-evaluate litter size data in rabbits and pigs under transformed scales.

Main Methods:

• Application of genetically structured heterogeneous variance models.
• Incorporation of the Box-Cox transformation family for data scaling.
• Re-analysis of rabbit and pig litter size data using transformed scales.

Main Results:

• Inferences on genetic components of environmental variance are sensitive to the scale of measurement.
• Rabbit data showed weaker evidence for genetic effects on environmental variance after transformation.
• Pig data showed stronger evidence, but the correlation between genetic effects on mean and variance changed sign.

Conclusions:

• Data asymmetry significantly affects inferences about genetic influences on variance.
• The choice of data scale is critical for accurate genetic variance component analysis.
• Future studies should validate normality assumptions or use appropriate transformations to avoid spurious genetic inferences.