Hydra

Hydra detects structural variation (SV) breakpoints by clustering discordant paired-end alignments that share corroborative signatures to enable multisample identification of SVs across unique and duplicated genomic regions.

Key Features:

• Comprehensive SV detection: Detects breakpoints across all classes of structural variation, including events in unique and duplicated sequences, by examining paired-end reads with multiple discordant alignments.
• Multisample analysis capability: Co-analyzes SV signals across many samples while addressing library insert size variability and integrating SV alignment signals over long-range genomic rearrangements involving disjoint loci.
• Scalability and performance: Implemented in C++ to enable co-analysis of hundreds to thousands of human genomes on commodity hardware.
• Accuracy and speed: Validated for accuracy, speed, and scalability using datasets from the 1000 Genomes Project and The Cancer Genome Atlas (TCGA).

Scientific Applications:

• Population genomics: Joint discovery and characterization of SVs across many human genomes to study allele frequency and population structure.
• Cancer genomics: Detection and analysis of somatic and germline structural variants across tumor and normal cohorts.
• Evolutionary studies: Identification of structural variants relevant to genome evolution and comparative analyses across samples.
• Variant discovery for complex traits: Joint SV calling across cohorts to support identification of variants that may underlie complex traits and diseases.

Methodology:

Hydra clusters discordant paired-end alignments by shared breakpoint signatures, examines reads with multiple discordant alignments to detect SVs in unique and duplicated regions, and integrates alignment signals across long-range rearrangements while accounting for library insert size variability.