lme4GS: An R-Package for Genomic Selection.

Area of Science:

• Quantitative genetics
• Statistical genomics
• Bioinformatics

Background:

• Genomic selection (GS) is a powerful tool for genetic improvement, surpassing traditional phenotype-based selection.
• Linear mixed models (LMMs) are crucial for addressing statistical challenges in GS and enabling genome-based predictions.
• Existing R packages like lme4 have limitations in defining complex covariance structures for genetic analyses.

Purpose of the Study:

• Introduce the novel lme4GS package for R, designed to enhance LMM fitting for genomic prediction.
• Address the limitations of current software in handling user-defined covariance structures for GS.
• Provide a flexible tool for advanced genomic prediction models in R.

Main Methods:

• Development of the lme4GS package in R, specifically tailored for LMMs in GS.
• Implementation of user-defined covariance structures and bandwidth selection within the package.
• Application of lme4GS to fit various GS models and analyze real-world genetic data.

Main Results:

• The lme4GS package effectively fits LMMs with user-specified covariance structures for genomic prediction.
• Demonstrated flexibility in handling different variance-covariance matrices for diverse GS models.
• Successful application of the package using real genetic datasets, validating its utility.

Conclusions:

• The lme4GS package offers a significant advancement for researchers in genomic selection and prediction.
• Its ability to define custom covariance structures overcomes key limitations of existing R packages.
• lme4GS provides a robust and flexible platform for advanced genetic analyses and breeding programs.