Estimation of heritability by parent-offspring regression

Area of Science:

• Quantitative genetics
• Plant breeding
• Genetics

Background:

• Narrow-sense heritability (h²) estimates are crucial for breeding programs.
• Previous inbreeding and genotype x environment (GxE) interactions can introduce bias into these estimates.
• Parent-offspring regression is a common method for estimating heritability.

Purpose of the Study:

• To enumerate the implications of bias from parental inbreeding and GxE interaction on heritability estimates.
• To propose methods for mitigating bias in heritability estimates.

Main Methods:

• Review and enumeration of bias sources in heritability estimation.
• Application of analysis of covariance (ANCOVA) models to correct for GxE interaction.
• Consideration of ANCOVA with heterogeneity of slopes for interacting organisms.

Main Results:

• Parental inbreeding and GxE interactions can significantly bias heritability estimates derived from parent-offspring regression.
• ANCOVA models can effectively remove bias caused by GxE interactions.
• In cases of significant host genotype x strain interactions (e.g., symbiotic N2 fixation), separate heritability estimates per strain are recommended.

Conclusions:

• Accurate heritability estimates require accounting for parental inbreeding and GxE interactions.
• ANCOVA provides a robust statistical framework for bias correction in heritability studies.
• Tailored heritability estimation strategies are necessary for complex biological interactions to maximize breeding efficiency.