GRIDSS

GRIDSS detects and localizes structural variants and genomic breakpoints from next-generation sequencing (NGS) data to characterize genomic rearrangements in cancer and other genomes.

Key Features:

• Positional de Bruijn graph break-end assembly: Uses a positional de Bruijn graph assembler to perform genome-wide break-end assembly prior to variant calling.
• Probabilistic scoring system: Integrates assembly, split reads, and read pairs with a probabilistic scoring framework to combine multiple lines of evidence for each call.
• High sensitivity and specificity: Demonstrates improved sensitivity and reduced false discovery rate on simulated data, cell line experiments, and patient tumor samples compared to contemporary methods.
• Comprehensive SV detection: Identifies a wide range of structural variants including non-template sequence insertions, micro-homologies, and large imperfect homologies.
• Multi-sample analysis support: Supports simultaneous analysis of multiple samples to facilitate comparative and large-scale genomic studies.

Scientific Applications:

• Cancer genomics: Detects structural variants and complex rearrangements relevant to tumorigenesis and tumor progression in cancer samples.
• Benchmarking and method evaluation: Serves as a validated approach in challenges such as the ICGC-TCGA DREAM Somatic Mutation Calling Challenge.
• Comparative and large-scale studies: Enables comparative analysis across multiple samples for large-scale genomic projects.

Methodology:

Performs break-end assembly using a positional de Bruijn graph assembler, integrates assembly, split-read, and read-pair evidence via a probabilistic scoring system, and calls structural variants from the integrated evidence.