HAST

HAST generates haplotype-resolved assemblies of diploid genomes by partitioning synthetic long-read (stLFR) cobarcoded second-generation sequencing reads with a trio-binning strategy to produce maternal and paternal haplotype assemblies.

Key Features:

• Trio-binning classification: Classifies sequencing reads into maternal and paternal bins using parental information.
• Parentally unique markers: Uses parentally unique markers to partition synthetic long-read (stLFR) cobarcoded reads.
• stLFR cobarcoded read support: Operates on cobarcoded second-generation sequencing data generated as synthetic long reads (stLFR).
• Independent haplotype assembly: Performs independent de novo assembly of each parent-specific haplotype after read partitioning.
• Assembly coverage: Produced haplotype assemblies covering 94.7% of the reference genome in an Asian individual dataset.
• Assembly contiguity: Assembled scaffolds with an N50 length exceeding 11 megabases (Mb).
• Precision and recall: Achieved approximately 99.7% precision and ~95.9% recall in reported benchmarks.
• Single-base accuracy: Reported single-base accuracy up to 99.99997% (Q65), exceeding that reported for TrioCanu on third-generation long-read data.
• Comparative benchmarking: Demonstrated higher precision and recall than Supernova for assembling cobarcoded reads.

Scientific Applications:

• Haplotype-resolved genome assembly: Generation of maternal and paternal haplotype assemblies for diploid genomes.
• Trio-based haplotyping: High-confidence haplotyping within human trios using parental sequencing information.
• Structural variant analysis: Investigation of structural variants in trio-based studies.
• Variant phasing for phenotype association: Linking the order of genetic variations along homologous chromosomes to phenotypic traits.
• Cobarcoded dataset analysis: Assembly and analysis of cobarcoded stLFR second-generation sequencing datasets.

Methodology:

Classifies reads by parental origin using a trio-binning algorithm with parentally unique markers to partition stLFR cobarcoded reads, followed by independent de novo assembly of each parent-specific haplotype.