An increased number of heterozygous calls in the AxiomTM Equine Genotyping Array.

An analysis of the Axiom™ Equine 670K SNP genotyping array revealed that some horses show an unusually high number of heterozygous calls, which may reflect issues with genotype accuracy and affect genetic analyses. The study evaluated genotype concordance across different testing batches and platforms, identifying best practices for reliable use of this SNP array in equine genetics.

Background and Purpose of the Study

• SNP arrays, such as the Axiom™ Equine 670K, are widely employed in livestock genetics to study complex traits, including genome-wide association studies (GWAS), fine mapping, genomic prediction, and genetic diversity analysis.
• In this research, genotype data from 2,768 equids spanning 20 horse breeds and one donkey breed were analyzed with a focus on runs of homozygosity (ROH) and genotype accuracy.
• The study aimed to understand why some horses display fewer ROH segments than expected and to evaluate genotype concordance to ensure data reliability.

Analysis of Runs of Homozygosity (ROH)

• ROH refers to continuous stretches of homozygous genotypes, which can indicate inbreeding or shared ancestry.
• Using a strict detection setting, purebred horses exhibited fewer ROH segments compared to F1 crosses, which was unexpected as purebreds are generally more inbred.
• Under more permissive (medium and relaxed) detection thresholds, the number of ROH segments identified increased, but some horses still showed abnormally low ROH counts.
• This anomaly prompted further examination of genotype data quality and reproducibility across platforms and batches.

Genotype Concordance and Data Quality Assessments

• Four-fold concordance analyses were performed comparing:
• Replicates genotyped on the same Axiom™ batch.
• Replicates genotyped on different Axiom™ batches.
• Genotypes obtained from Illumina EquineSNP50 BeadChip®.
• Genotypes derived from Illumina paired-end HiSeq 2000 whole-genome sequencing data.
• Concordance rates indicated:
• Highest agreement (98.81%) was seen between replicates from the same Axiom™ batch, signifying high internal consistency.
• Concordance decreased slightly with different platforms: Illumina 50K (97.88%) and whole-genome sequencing (96.84%).
• Horses with few ROH segments had the lowest concordance rate (93.52%) upon re-genotyping, suggesting possible genotyping errors or sample issues.
• Using SNPolisher™ software, 120,838 SNPs across the genome were identified as “not recommended” due to low reproducibility and potential quality issues.

Recommendations for Improved Genotyping Analysis

• Combining genotype calls from different batches using Axiom™ Best Practice guidelines (including removal of failing samples prior to final genotyping) improved concordance across datasets.
• Excluding horses with an unusually high number of heterozygous calls is advisable because such data likely reflect genotype calling errors.
• Utilizing only SNP markers validated for genotype performance enhances data reliability.
• Accounting for batch effects is important when analyzing data derived from multiple genotyping batches to prevent spurious results or bias.

Implications for Equine Genetics Research

• The presence of increased heterozygous calls and low ROH counts in some purebred horses suggests technical artifacts rather than biological differences.
• Careful quality control, including SNP and sample filtering, is critical to ensure valid downstream analyses such as GWAS or population studies.
• This study provides practical guidelines to improve the reproducibility and accuracy of genotyping data from the Axiom™ Equine 670K SNP array, thereby supporting more robust genetic research in equines.