Advancements in Animal Breeding: From Mendelian Genetics to Machine Learning

Area of Science:

• Animal Genetics
• Genomics
• Machine Learning in Animal Breeding

Background:

• Livestock breeding evolved from Neolithic domestication and Mendelian genetics.
• Quantitative genetics advanced the study of complex traits like body weight and milk yield.
• Genomic selection (GS) and genomic best linear unbiased prediction (GBLUP) improved breeding accuracy.

Purpose of the Study:

• To review the historical progression of animal breeding techniques.
• To highlight the integration of genomic and machine learning (ML) approaches.
• To showcase advancements in livestock productivity, health, and welfare.

Main Methods:

• Exploration of historical breeding paradigms from phenotypic to genotypic selection.
• Review of quantitative genetics, QTL mapping, and genomic prediction methodologies.
• Analysis of machine learning and artificial intelligence applications in genomic prediction.

Main Results:

• QTL mapping enabled genotype-based selection, improving accuracy.
• Genomic estimated breeding values (GEBV) under GBLUP were enhanced by QTL associations.
• Deep learning models achieved high accuracy (0.643 correlation) in genomic prediction.
• Multi-omics strategies advanced disease management in livestock.

Conclusions:

• Innovative methods, particularly ML, are transforming animal breeding.
• These advancements significantly impact milk and meat production and disease management.
• Future prospects focus on enhancing livestock productivity, health, and welfare through integrated approaches.