Machine Learning-Aided Ultra-Low-Density Single Nucleotide Polymorphism Panel Helps to Identify the Tharparkar Cattle...

Area of Science:

• Livestock Genomics
• Artificial Intelligence in Agriculture
• Animal Genetics

Background:

• Cattle breed identification is vital for livestock research and sustainable food systems.
• Genomics and artificial intelligence (AI) offer advanced tools for livestock management.
• The Tharparkar cattle breed requires specific identification methods for conservation and research.

Purpose of the Study:

• To investigate the identification of the Tharparkar cattle breed using genomics and machine learning (ML).
• To develop a robust, breed-specific panel of single nucleotide polymorphisms (SNPs) for Tharparkar cattle.
• To determine the minimal number of SNPs required for accurate breed identification.

Main Methods:

• Leveraged data from the Bovine SNP 50K chip to develop a breed-specific SNP panel.
• Integrated data from seven other Indian cattle populations to enhance panel robustness.
• Employed genome-wide association studies (GWAS), principal component analysis, and ML models (AdaBoost, bagging tree, gradient boosting machines, random forest) to refine SNP selection.

Main Results:

• Identified a panel of 500 SNPs, subsequently refined to minimal sets of 23 and 48 SNPs.
• Achieved high accuracy rates of 95.2-98.4% for Tharparkar cattle breed identification using these minimal SNP panels.
• The identified SNPs are associated with important productive and adaptive traits in cattle.

Conclusions:

• The ML-aided ultra-low-density SNP panel approach is effective for accurate cattle breed identification.
• This method facilitates the preservation of genetic diversity and supports future cattle breeding programs.
• Highlights the potential of digital transformation in advancing livestock genomics and sustainable agriculture.