ETCHING

ETCHING detects somatic structural variations (SVs) by filtering reads that match pan-genome and normal k-mer sets to reduce mapping costs and improve SV detection accuracy in cancer genomics.

Key Features:

• Read filtering (pan-genome and normal k-mer sets): Filters reads that match both pan-genome and normal k-mer sets prior to mapping to reduce computational mapping costs.
• Random Forest classifier: Applies a Random Forest classifier that uses six breakend-related features to minimize false positives in called somatic SVs.
• SV detection modes: Detects somatic SVs from tumor and matched-normal whole genomes and from tumor-only targeted sequencing datasets.
• Benchmarking and validation: Has been benchmarked using reference SV materials and validated SV biomarkers across multiple datasets.
• Computational performance: Reports at least 15-fold faster runtime than competing tools while maintaining memory efficiency.

Scientific Applications:

• Cancer genomics: Identification of somatic structural variations in tumor genomes for research and clinical studies.
• Biomarker validation: Assessment and validation of known SV biomarkers and reference SV materials.
• Large-scale genomics: Enabling large-scale whole-genome and targeted sequencing projects by reducing mapping and computational costs.
• Precision medicine: Rapid SV detection to support clinical and real-time precision-medicine analyses.

Methodology:

Reads matching pan-genome and normal k-mer sets are filtered prior to mapping to reduce mapping cost, and candidate breakpoints are scored by a Random Forest classifier using six breakend-related features to reduce false positives.