KAML

KAML enhances genomic prediction accuracy for complex traits by combining kinship-adjusted multiple-loci Best Linear Unbiased Prediction (BLUP) with machine-learning-driven parameter optimization.

Key Features:

• Machine Learning Integration: Employs a machine-learning-based approach that integrates cross-validation, multiple regression, grid search, and bisection algorithms to optimize predictive parameters.
• Efficiency in Computation: Draws on efficient computation approaches for linear mixed models and concepts from Bayesian methods to balance prediction accuracy with computational efficiency.
• Improved Prediction Accuracy: Demonstrates improved prediction accuracy compared to existing genomic prediction methods for complex traits.

Scientific Applications:

• Animal and Plant Breeding: Provides more precise genetic evaluations to improve selection decisions in breeding programs.
• Human Genetics: Supplies improved predictive models for studying complex human traits and diseases.

Methodology:

KAML combines kinship-adjusted multiple-loci BLUP with a machine-learning framework using cross-validation, multiple regression, grid search, and bisection algorithms and is informed by methods for linear mixed models and Bayesian prediction.