Identification of Chicken Populations by Machine Learning Models Using the Minimum Number of SNPs

Dongwon Seo, Sunghyun Cho, Prabuddha Manjula, Nuri Choi, Young‐Kuk Kim, Yeong Jun Koh, Seung Hwan Lee, Hyung-Yong Kim, Jun Heon LeeResearch 2020 · DOI

Abstract

Verifying the origin and breed of chicken populations is essential for maintaining commercial value and protecting genetic resources. This study presents a streamlined pipeline that identifies specific chicken populations using a minimal set of genetic markers. By integrating genome-wide association studies (GWAS), linkage disequilibrium (LD) analysis, and machine learning (ML) algorithms, researchers reduced the required markers from 600,000 SNPs to as few as 8–44 SNPs. Models such as AdaBoost and Random Forest achieved classification accuracies exceeding 97%, demonstrating that high-precision breed identification is possible with significantly reduced genomic data.

Introduction

In the poultry industry, the ability to verify a bird's breed or origin is critical for consumer confidence and the preservation of unique genetic lines, such as the Korean Native Chicken (KNC). Traditional morphological identification is ineffective once products reach the market, and existing genetic methods—like microsatellite markers or high-density SNP arrays—are often too labor-intensive or expensive for routine commercial use. This research addresses the need for a "Goldilocks" set of markers: a combination small enough for cost-effective testing yet powerful enough to distinguish closely related populations.

Methods

The study utilized a multi-stage computational pipeline to filter 600,000 SNPs down to a functional minimum:

  1. High-Density Genotyping: Initial data was gathered using a 600K SNP array across 283 samples from 20 chicken lines.
  2. Statistical Filtering: GWAS identified SNPs significantly associated with the target "case" group versus "control" groups.
  3. LD Pruning: To eliminate redundancy, LD analysis ensured selected markers were distributed across the genome rather than clustered.
  4. Machine Learning & Feature Selection: Eight ML models were evaluated. The "feature selection" capabilities of these models identified the absolute minimum number of SNPs required for high predictive power.

Performance was evaluated using standard classification metrics, including the True Positive Rate (TPR) and True Negative Rate (TNR), to ensure both sensitivity and specificity in breed identification.

True Positive Rate (TPR) formula used to evaluate model sensitivity.
True Negative Rate (TNR) formula used to evaluate model specificity.

Results and Discussion

The researchers identified three optimized marker sets tailored for different machine learning models, each maintaining high accuracy despite the drastic reduction in data:

  • AdaBoost (AB): Achieved 99.6% accuracy using 36 SNPs.
  • Random Forest (RF): Achieved 98.0% accuracy using 44 SNPs.
  • Decision Tree (DT): Achieved 97.9% accuracy using just 8 SNPs.

Key Insights

  • Pruning Redundancy: Selecting SNPs based solely on statistical significance (p-values) often resulted in poor clustering because those SNPs were often in high LD (physically close on the chromosome). Pruning for LD allowed for a more representative genomic "fingerprint."
  • Efficiency of Scale: Near-perfect accuracy was achieved with fewer than 45 markers, a significant reduction from the hundreds suggested in previous literature.
  • Genetic Distance: The minimal marker sets actually increased the calculated genetic distance between target and control groups by filtering out "genetic noise" that typically obscures breed identity.

Conclusion

The combination of GWAS, PCA, and machine learning provides an efficient framework for identifying the minimum marker sets necessary for population classification. While the Decision Tree model used the fewest markers, AdaBoost and Random Forest showed greater robustness in validation sets. Future work will focus on the physical development of low-cost, automated SNP kits based on these markers for use in commercial food safety and traceability labs.

Summary updated

Abstract

Abstract BackgroundA marker combination capable of classifying a specific chicken population could improve commercial value by increasing consumer confidence with respect to the origin of the population. This would also facilitate the protection of genetic resources, especially in developing countries. MethodsIn this study, a total of 20 lines 283 samples which were consist of Korean native chicken, commercial native chicken, and commercial broilers with layer population were used for finding the minimum number of marker combinations through the 600k high-density single nucleotide polymorphism (SNP) array. Application of the machine learning algorithms, a genome-wide association study (GWAS), linkage disequilibrium (LD) analysis, and principal component analysis (PCA) were used to distinguish a target (case) group from control chicken groups. In the verification of the selected markers, a total of 12 lines 182 samples were used to confirm the change in the accuracy of the target chicken breed identification.ResultsA total of 47,303 SNPs was used for classifying chicken populations; 96 LD-pruned SNPs (50 SNPs per LD block) served as the best marker combination for target chicken classification. Moreover, 36, 44, and 8 SNPs were selected as the minimum numbers of markers by Adaboost (AB), Random Forest (RF), and Decision Tree (DT) machine learning classification models, which had accuracy rates of 99.6%, 98.0% and 97.9%, respectively. The selected marker combinations increased the genetic distance between the case and control groups, and reduced the number of genetic components, confirming that an efficient classification of the groups was possible using small number of marker sets. In a verification study including additional chicken breeds and samples, the accuracy did not significantly change, and the target chicken group could be clearly distinguished from the other populations.ConclusionsThe GWAS and PCA analysis, machine learning algorithm used in this study is able to be applied efficiently to explore the minimum combination of markers that can distinguish varieties among a large number of SNP markers.
BibTeX
@inproceedings{seo2020identification,
  author = {Dongwon Seo and Sunghyun Cho and Prabuddha Manjula and Nuri Choi and Young‐Kuk Kim and Yeong Jun Koh and Seung Hwan Lee and Hyung-Yong Kim and Jun Heon Lee},
  title = {{Identification of Chicken Populations by Machine Learning Models Using the Minimum Number of SNPs}},
  booktitle = {Research},
  year = {2020},
  doi = {10.21203/rs.3.rs-95706/v1},
}

Loading related content...

Figure