Practical Haplotype Graph

Practical Haplotype Graph represents pangenomes as a trellis graph anchored to a reference genome and enables imputation of complete genomes from low-density sequence or variant data for genomic diversity analysis.

Key Features:

• Trellis Graph Representation: Employs a trellis graph structure anchored to a reference genome to model pangenome variation.
• Imputation from Low-density Data: Enables reconstruction and imputation of complete genomes and haplotypes from low-density sequence or variant data.
• Data Compression and Storage: Achieves up to 1000-fold data compression and uses PostGRES and SQLite databases for data management.
• Imputation Accuracy: Demonstrates accurate haplotype reconstruction on simulated datasets at coverages as low as 0.1×.
• Programming Flexibility: Provides components and implementations in Java, Kotlin, and R.
• Integration with Breeding API (BrAPI): Integrates with BrAPI to exchange genomic data relevant to breeding applications.
• Application in Crop Genomics: Applied to represent genetic diversity in maize and other major crops with complex pangenomes.

Scientific Applications:

• Population Genetics: Enables analysis of population-level haplotype diversity using a pangenome-aware representation and imputation.
• Quantitative Genetics: Supports genotype imputation and haplotype-based analyses for quantitative trait mapping.
• Breeding: Facilitates genomic selection and breeding workflows by providing imputed genotypes and interoperability with BrAPI.
• Low-coverage Sequencing Studies: Allows genomic analyses and haplotype inference from sparse sequencing datasets to reduce sequencing requirements.

Methodology:

Uses a trellis graph representation anchored to a reference genome, stores haplotypes in PostGRES and SQLite, and performs haplotype imputation evaluated on simulated datasets down to 0.1× coverage.