Hierarchical mixed-model expedites genome-wide longitudinal association analysis

Area of Science:

• Quantitative genetics
• Genomic data analysis
• Statistical modeling

Background:

• Longitudinal data analysis in genomics presents challenges due to high dimensionality.
• Existing random regression models (RRM) can be computationally intensive for genome-wide association studies (GWAS).
• Efficiently linking phenotypic trajectories to genetic markers is crucial for understanding complex traits.

Purpose of the Study:

• To extend the hierarchical random regression model (Hi-RRM) for genome-wide association analysis of longitudinal data.
• To reduce the dimensionality of repeated measurements in genetic analyses.
• To develop a computationally efficient method for associating phenotypic regressions with genetic markers.

Main Methods:

• The hierarchical random regression model (Hi-RRM) was employed, modeling individual phenotypic trajectories with RRM.
• A multivariate mixed model (mvLMM) associated phenotypic regressions with genetic markers.
• Spectral decomposition of genomic relationship and regression covariance matrices transformed mvLMM into multiple linear regression.
• Implementation of mvLMM associations was achieved using efficient mixed-model association expedited (EMMAX).

Main Results:

• The Hi-RRM significantly reduced the dimensionality of longitudinal data.
• The mvLMM transformation improved computing efficiency compared to existing RRM-based methods.
• Simulation experiments demonstrated the statistical utility of Hi-RRM.
• The method successfully identified quantitative trait nucleotides controlling egg weight growth patterns in poultry.

Conclusions:

• The developed Hi-RRM provides a statistically robust and computationally efficient approach for GWAS on longitudinal data.
• This method enhances the ability to identify genetic factors influencing growth patterns and other time-dependent traits.
• The application to poultry data highlights the practical utility of Hi-RRM in quantitative genetics research.