PASHA

PASHA assembles large genomes from short-read sequencing data using de Bruijn graph methods and parallel computation to enable scalable de novo assembly.

Key Features:

• De Bruijn graph-based assembly: Implements de Bruijn graph methodologies for constructing genome assemblies from short reads.
• Parallelization: Employs parallel computing techniques to improve throughput and resource utilization.
• Hybrid computing architecture: Supports combined shared-memory multi-core CPU and distributed-memory compute cluster execution.
• Scalability for large genomes: Optimized for large genome datasets, including human genome assemblies.
• Support for paired-end data: Demonstrated performance on real paired-end short-read datasets.

Scientific Applications:

• High-quality de novo assembly: Produces assemblies with greater contiguity compared to Velvet, ABySS, and SOAPdenovo on tested datasets.
• Performance metrics: Reported results include an NG50 contig size of 503, longest correct contig size of 18,252, and NG50 scaffold size of 2,294.
• Human genome assembly: Capable of completing a human genome assembly in approximately 21 hours on modest compute resources.
• Adaptation to growing NGS datasets: Parallelized approach enables handling of increasing next-generation sequencing data volumes.

Methodology:

Uses de Bruijn graph-based assembly combined with parallel computing on hybrid shared-memory multi-core CPUs and distributed-memory compute clusters.