HISEA

HISEA performs sensitive alignment of long Single Molecule, Real-Time (SMRT) sequencing reads from Pacific Biosciences to detect read overlaps for genome assembly while accommodating high indel error rates (~10–15%).

Key Features:

• Two-step k-mer search: Employs a two-step k-mer search strategy comprising consistent clustering, k-mer filtering, and read alignment extension to improve alignment accuracy for long SMRT reads.
• Comparative performance: Demonstrated superior sensitivity relative to BLASR, DALIGNER, GraphMap, MHAP, and Minimap across precision, specificity, F1-score, runtime, and memory usage.
• Evaluation methodology: Introduced a new evaluation method that provides more precise assessments of alignment performance.
• Integration with Canu: Integrated into the Canu genome assembly pipeline and shown to improve assembly quality compared with MHAP within Canu.
• Improved assembly outcomes: Enabled better assemblies in benchmarks at 30× and 50× coverage, with Canu+HISEA at 30× comparable to Canu+MHAP at 50× while being faster and more cost-effective.

Scientific Applications:

• SMRT-based genome assembly: Enables higher-quality genome assemblies from Pacific Biosciences SMRT long-read data by improving overlap detection.
• Assembly of complex genomes: Facilitates assembly of genomes with repetitive or complex regions that challenge short-read technologies by handling long reads with high indel rates.
• Coverage and cost reduction: Supports reduced sequencing coverage requirements by achieving comparable assembly quality at lower coverage (e.g., 30× vs 50×), potentially lowering costs.

Methodology:

Uses a two-step k-mer search (consistent clustering, k-mer filtering, read alignment extension), applies a newly introduced evaluation method for alignment assessment, and was benchmarked against BLASR, DALIGNER, GraphMap, MHAP, and Minimap.