GRIDSS: sensitive and specific genomic rearrangement detection using positional de Bruijn graph assembly.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Accurate identification of genomic rearrangements is crucial for precision medicine and cancer research.
• Massively parallel sequencing presents challenges in detecting these structural variants (SVs) with high sensitivity and specificity.

Purpose of the Study:

• To introduce GRIDSS (Genome Rearrangement IDentification Software Suite), a novel method for detecting genomic rearrangements.
• To demonstrate GRIDSS's high sensitivity and specificity in identifying structural variants.

Main Methods:

• GRIDSS employs a multithreaded approach for efficient genome-wide break-end assembly using a positional de Bruijn graph-based assembler.
• It combines evidence from assembly, split reads, and read pairs with a probabilistic scoring system.
• The software supports multisample analysis and provides quality scores for breakpoints.

Main Results:

• GRIDSS achieved high sensitivity and specificity on simulated, cell line, and patient tumor data.
• It won the SV subchallenge #5 of the ICGC-TCGA DREAM8.5 Somatic Mutation Calling Challenge.
• On human cell line data, GRIDSS reduced the false discovery rate by half while maintaining or improving sensitivity compared to other methods.

Conclusions:

• GRIDSS offers a significant advancement in the accurate detection of genomic rearrangements.
• The software is capable of identifying various types of structural variants, including insertions and homologies.
• GRIDSS provides a robust tool for cancer research and precision medicine applications.