PHARP

PHARP provides genotype phasing and imputation for pig genomics using a reference panel of 2,012 haplotypes across approximately 34 million single nucleotide polymorphisms (SNPs) derived from whole-genome sequencing to enable high-density genotype inference from low-density SNP data.

Key Features:

• Extensive haplotype reference: A reference panel containing 2,012 pig haplotypes at approximately 34 million SNPs constructed from whole-genome sequence data sourced from over 49 studies spanning 71 pig breeds.
• Phasing and imputation: Implements phasing and genotype imputation for pig datasets to infer high-density genotypes from lower-density data.
• Concordance performance: Demonstrated a concordance rate of 0.971 when imputing commercial 50K SNP array data for Large White pigs with average sequencing coverage of ~17×.
• Downstream analysis support: Produces imputed genotypes usable for GWAS, TWAS, eQTL mapping, and genomic prediction (GS), facilitating detection of significantly associated SNPs potentially linked to causal variants.
• Low-density SNP imputation application: Applied to impute low-density SNP chip data into high-density datasets and used in three GWASs where it identified novel significant associations.

Scientific Applications:

• Genome-wide association studies (GWAS): Enables higher-resolution association mapping by imputing additional SNPs for GWAS analyses.
• Transcriptome-wide association studies (TWAS): Supports TWAS by providing denser genotype data for integration with transcriptomic data.
• eQTL mapping: Facilitates expression quantitative trait loci mapping through increased genotype density.
• Genomic prediction and selection (GS): Improves genomic prediction accuracy for breeding programs by supplying high-density genotype calls.
• Disease-resistance and comparative genomics: Supports studies of disease resistance and comparative genomics between pigs and humans through accurate imputation and phasing.

Methodology:

Reference panel assembled from whole-genome sequencing data from over 49 studies covering 71 breeds to generate 2,012 haplotypes at ~34 million SNPs; phasing and genotype imputation were applied to low-density SNP data and performance was evaluated by concordance (0.971) against commercial 50K SNP arrays for Large White pigs with ~17× sequencing coverage.