MegaLMM

MegaLMM implements a linear mixed-model framework for genomic prediction that jointly models thousands of phenotypic traits to improve genetic value estimation.

Key Features:

• Scalability: Manages and analyzes thousands of traits simultaneously, overcoming computational limits of traditional multivariate linear mixed effect models.
• Enhanced Prediction Accuracy: Leverages extensive phenotype datasets to improve accuracy of genetic value predictions, demonstrated on plant datasets.
• Robust Statistical Framework: Employs a statistical framework that addresses limitations of traditional multivariate linear mixed effect models for reliable multi-trait genomic prediction.

Scientific Applications:

• Plant Breeding: Improves selection and genetic improvement strategies in plant breeding by incorporating large-scale phenotype data.
• Animal Genetics: Enhances genomic predictions for livestock and other animal species to inform breeding decisions and trait management.
• Human Genetics: Enables simultaneous analysis of numerous traits in human genetics to study complex genetic interactions and disease associations.

Methodology:

Employs linear mixed models tailored for high-dimensional data and integrates phenotype datasets with genomic information to jointly model relationships among multiple traits.